Quantitative microbial risk assessment for sustainable water resources
A Thesis
Submitted to the Faculty
of
Drexel University
by
Kerry Ann Hamilton
in partial fulfillment of the
requirements for the degree
of
Doctor of Philosophy
December 2016
© Copyright 2016
Kerry Ann Hamilton. All Rights Reserved.
iii
Dedication
For Nick, my parents, and my brother Mark.
iv
Acknowledgement
First I would like to thank my advisor, Dr. Charles Haas for encouraging me on my path to pursue an engineering degree and allowing me tremendous freedom to pursue projects
I am passionate about at the nexus of environmental engineering and public health. He has been extremely patient and provided many opportunities for me to grow as a researcher and person over the past five years. I am extraordinarily grateful for his expert mentorship and support during this time.
This work was greatly enriched by my committee members Dr. Patrick Gurian, Dr.
Mira Olson, Dr. Franco Montalto, Dr. Mark Weir, and Dr. Christopher Sales. Their expertise has greatly improved the methodology and insights of this work, and they have provided generous feedback throughout this journey. I would like to extend a special thanks to the
Australian-American Fulbright Commission and my Fulbright host laboratory at the
Australian Commonwealth Scientific and Industrial Research Organization (CSIRO).
Additionally, I am especially grateful for my mentors Dr. Simon Toze and Dr. Warish Ahmed for giving me many opportunities, intensive laboratory training, and leveraging funding support to conduct a large study of roof-havested rainwater tanks as a visiting member of their research group. Additionally, CSIRO group members Leonie Hodgers, Andrew Palmer, Kylie
Smith, Dr. Pradip Gyawali, Dr. Jatinder Sidhu, and numerous other CSIRO staff have aided significantly in both logistical and technical aspects of this work and I am greatly appreciative of their efforts. I am also very thankful to numerous employees of the Ecosciences Precinct in
Brisbane, Australia as well as residents at the Currumbin Ecovillage for generously allowing me to sample their rainwater tanks. Their insightful questions and feedback improved the survey methodology and interpretation of the results greatly and this work could not have been performed without their voluntary participation. Tremendous thanks are due to the Haas lab collaborators on the Legionella risk management project at American Water, Dr. Mark
LeChevallier, Dr. Patrick Jjemba, and Mr. William Johnson for their microbiological insights, patience, and help in developing our risk assessment model.
v
I would like to thank my fellow researchers in the Haas lab group current and former:
Bidya Prasad, Dr. Michael Ryan, Dr. Neha Sunger, and Dr. Sandra Teske. Specifically I would like to thank Dr. Ryan for introducing me to the process of risk modelling and helping to trouble-shoot models and laboratory experiments throughout my time at Drexel and Bidya
Prasad for her modeling expertise and extraordinary caring and willingness to help with any project. I would also like to thank my fellow graduate students current and former including
Kaitie Sniffen, Anita Avery, Dr. Kimberley DiGiovanni, Stephen White, Dr. Kimberlee
Marcellus, Dr. Noura Abualfaraj, Dr. Megan Hums, Doug Goetz, Bita Alizadehtazi, Dr.
Somayeh Yoosefi, Dr. Qasideh Pourhashem, Dr. Raquel Catalano de Sousa, and Yetunde
Sorunmu for their support and encouragement. Thank you to Drexel Graduate Women in
Science and Engineering founder Dr. Josa Hanzlik and board members Deeksha Seth, Dr.
Marissa Powers, Val Tutwiler, and Dr. Kristyn Voegele for their leadership and support to create opportunities for graduate students and greatly enriching the graduate experience for myself and others. Thank you to graduate and administrative staff Barbara Interlandi,
Kenneth Holmes, Sarah Collins, Sharon Stokes, Kim Spina, Teck-Kah Lim, and especially
Taz Kwok for always coming to the rescue many times throughout the past five years and streamlining the graduate school process. Thank you to Dr. Robert Brehm for encouraging me to take the F.E. exam and helping me throughout the preparation process. Thank you to Jay
Bhatt and the library staff for their help during our extensive literature review process and numerous article requests.
I would like to acknowledge the funding opportunities provided to me throughout this research from WateReuse Research Foundation grant WRF-05 and the Fulbright-CSIRO
Postgraduate Scholarship from the Australian-American Fulbright Program. Additionally, I have received generous support from Drexel University through the Drexel Provost Graduate
Fellowship, the Drexel George Hill Jr. Fellowship, the Steven E. Giegerich Memorial
Scholarship, the Drexel Higher Education Advocate Travel Award, the Marilyn A. Burshtin
Memorial Award, and the Claudio Elio Memorial Fellowship in Environmental Science &
vi
Environmental Engineering. The Society for Risk Analysis has supported my work through the dose response group annual meeting student merit award.
Finally, and most importantly, I would like to acknowledge my friends and family for their support. My husband, best friend, and sampling assistant Nicholas Preston has provided me with endless faith, patience, and support. My brilliant mom and dad have sacrificed many things in their lives to provide me with every opportunity to pursue my dreams. Thank you to my extraordinary brother Mark for inspiring me to always learn new things, and for his patience and technical coding ingenuity throughout the Legionella reclaimed water project.
Thank you also to Nanny, Aunt Lois, Uncle John, Michael, Danielle, Nicole, Steven, Aimee,
Michele, and Joe for your unwavering support and sense of humor. Cassandra and Alison thank you for helping me plan a wedding and continuing to be my best friends even throughout a very busy PhD program. Finally, thank you to mom and dad Preston, Anthony,
Vince, and Devon for your support.
To all of you—thank you will never be enough.
vii
Table of Contents
Table of Contents ...... vii List of Tables ...... xii List of Figures ...... xvi Abstract ...... xx 1. Introduction ...... 1 1.1. Background ...... 1 1.2. Research Objectives ...... 3 1.2.1. Objective 1: Prioritize pathogens previously quantified in harvested rainwater for further study ...... 3 1.2.2. Objective 2: Conduct a screening of Brisbane rainwater tanks for pathogens ...... 3 1.2.3. Objective 3: Conduct a seasonal study of Brisbane rainwater tanks for pathogens 4 1.2.4. Objective 4: Develop a dose response model for the Mycobacterium avium complex (MAC) ...... 4 1.2.5. Objective 5: Assess the health risks of Legionella and MAC in harvested rainwater 5 1.2.6. Objective 6: Assess the health risks of Legionella in reclaimed water ...... 5 2. Prioritization of pathogens in harvested rainwater ...... 7 2.1. Roof-harvested rainwater ...... 7 2.1.1. Introduction ...... 7 2.1.2. Human health Risks ...... 8 2.2. Pathogens identified for further study ...... 17 2.2.1. Legionella pneumophila ...... 17 3. Screening study of Brisbane roof-harvested rainwater tanks for opportunistic pathogens 25 3.1. Abstract ...... 25 3.2. Keywords: ...... 26 3.3. Introduction ...... 26 3.4. Materials and methods ...... 29 3.4.1. Study areas and survey ...... 29 3.4.2. Tank water sampling ...... 30 3.4.3. Enumeration of fecal indicator bacteria (FIB) ...... 30 3.4.4. Concentration of rainwater samples ...... 30 viii
3.4.5. DNA extraction ...... 31 3.4.6. PCR inhibition ...... 31 3.4.7. Preparation of qPCR standards ...... 31 3.4.8. qPCR assays ...... 32 3.4.9. Recovery efficiency ...... 33 3.4.10. Statistical analysis ...... 33 3.5. Results ...... 34 3.5.1. Survey data ...... 34 3.5.2. Fecal indicator bacteria (FIB) ...... 37 3.5.3. qPCR standards, lower limit of detection (LLOD) and quantification (LLOQ) 39 3.5.4. Concentrations of potential opportunistic pathogens in tank water samples ... 39 3.5.6. Correlations among FIB and opportunistic pathogens ...... 42 3.6. Discussion ...... 44 3.7. Conclusions ...... 48 4. Seasonal assessment of opportunistic pathogens in Brisbane roof-harvested rainwater tanks ...... 49 4.1. Abstract ...... 49 4.2. Keywords: ...... 50 4.3. Introduction ...... 50 4.4. Materials and Methods ...... 52 4.4.1. Tank water sampling...... 52 4.4.2. Tank survey ...... 53 4.4.3. Enumeration of FIB ...... 53 4.4.4. Concentration of rainwater samples and DNA extraction ...... 53 4.4.5. PCR inhibition ...... 54 4.4.6. qPCR standards ...... 54 4.4.7. qPCR assays and performance characteristics ...... 55 4.4.8. Quality control ...... 55 4.4.9. Meteorological data ...... 56 4.4.10. Statistical analysis ...... 56 4.4.11. Differences in binary (presence/absence) occurrence across six sampling events ...... 56 4.4.12. Differences in concentrations of FIB / OPPPs (continuous) occurrence across six sampling events ...... 57
ix
4.4.13. Correlations among FIB/OPPPs and between FIB/OPPPs and meteorological factors ...... 57 4.5. Results ...... 58 4.5.1. Survey data...... 58 4.5.2. Meteorological data ...... 61 4.5.3. FIB and OPPPs in tank water samples collected in phase 1 ...... 61 4.5.4. FIB and OPPPs in tank water samples collected in phase 2 ...... 62 4.5.5. Differences in FIB and OPPPs occurrence across six sampling events ...... 67 4.5.6. Correlations among FIB and OPPPs ...... 67 4.5.7. Correlations between FIB/OPPPs and meteorological factors ...... 71 4.6. Discussion ...... 72 5. Dose response models for Mycobacterium avium complex (MAC) ...... 78 5.5. Abstract: ...... 78 5.6. Keywords: ...... 79 5.7. Introduction ...... 79 5.8. Literature review ...... 85 5.8.6. Search strategy ...... 85 5.8.7. Curve fitting ...... 87 5.9. A framework for MAC quantitative microbial risk assessment ...... 89 5.9.6. MAC exposure routes ...... 89 5.9.7. MAC pathogenesis ...... 90 5.9.8. Predominating MAC species ...... 91 5.10. Dose response models for MAC ...... 95 5.10.6. Animal models ...... 95 5.11. Epidemiologic support for dose-response models ...... 107 5.11.6. Hypersensitivity pneumonitis outbreaks ...... 107 5.11.7. Other MAC infection and disease cases with quantitative dose information. 110 5.12. Incorporation of dose response models into the QMRA framework ...... 113 5.12.6. Selecting an appropriate dose response model for a particular population and health endpoint ...... 113 5.12.7. Extrapolating from experimental exposure routes ...... 114 5.12.8. Conditional probabilities for disease outcomes ...... 116 5.13. Limitations and research gaps ...... 119 5.13.6. Variation in virulence between MAC species ...... 119 5.13.7. Impact of environmental conditions on virulence ...... 120
x
5.13.8. Impact of microbial ecology on virulence ...... 121 5.13.9. Animal models ...... 122 5.14. Conclusions ...... 124 6. Quantitative microbial risk assessment (QMRA) of Legionella and MAC in roof- harvested rainwater ...... 126 6.1. Abstract ...... 126 6.2. Keywords ...... 127 6.3. Introduction ...... 127 6.4. Materials and Methods ...... 130 6.4.1. Exposure models ...... 130 6.4.2. Pathogen concentrations in RHRW ...... 141 6.4.3. Dose response ...... 142 6.4.4. Risk characterization ...... 142 6.5. Results ...... 151 6.5.1. Ingestion ...... 151 6.5.2. Inhalation ...... 154 6.5.3. Total annual risks ...... 159 6.6. Discussion ...... 162 6.7. Conclusions ...... 166 6.8. Acknowledgements ...... 166 7. Quantitative microbial risk assessment (QMRA) of Legionella in reclaimed water ..... 167 7.5. Introduction ...... 167 7.6. Hazard identification ...... 168 7.7. Methodology ...... 168 7.7.6. Exposure Models ...... 168 7.7.7. Legionella concentrations in reclaimed water and drinking water ...... 184 7.7.8. Decay rates ...... 189 7.7.9. Dose response and risk characterization ...... 192 7.8. Results and discussion ...... 194 7.8.6. Toilet flushing ...... 194 7.8.7. Cooling towers ...... 202 7.8.8. Spray irrigation ...... 210 7.9. Conclusions ...... 217 7.10. Acknowledgements ...... 222
xi
8. Conclusions ...... 223 8.1. Summary ...... 223 8.2. Future work ...... 224 9. Appendix ...... 226 9.1. RHRW literature review ...... 226 9.1.1. Factors that impact RHRW quality ...... 230 9.1.2. Impacts of RHRW use ...... 233 9.1.3. Research gaps ...... 241 9.2. Screening paper supplemental ...... 246 9.2.1. Supplemental figures and tables ...... 246 9.2.2. Recruitment of participants ...... 250 9.2.3. Survey administered to participants ...... 251 9.2.4. Dissemination of study results to participants ...... 257 9.3. Seasonal rainwater paper supplemental ...... 263 9.4. MAC literature review ...... 280 9.5. RHRW QMRA...... 297 9.6. A critical review of approaches for Legionella QMRA ...... 298 9.6.6. Abstract ...... 298 9.6.7. Introduction ...... 299 9.6.8. Frameworks for Legionella risk assessment ...... 301 9.6.9. Limitations of current models and research needs for QMRA development . 324 9.7. Legionella reclaimed water QMRA ...... 338 9.7.6. Cooling towers ...... 338 9.7.7. Sprinklers ...... 343 10. References ...... 347 11. Vita ...... 420
xii
List of Tables
Table 2.1 Best-fit dose response model parameters for pathogenic microorganisms reported in RHRW (qmrawiki.canr.msu.edu) ...... 13
Table 2.2 Ranges of microbial contaminants reported for RHRW across multiple studies (V= vegetated roofs, NV= non-vegetated roofs). Original references are provided in Chapter 9 Appendix Table 9.2...... 15
Table 2.3 Outbreaks/clusters of LD 2006-present (NA= Not Available) ...... 21
Table 3.1 Water uses reported by Currumbin (n = 45) and Brisbane (n = 76) rainwater survey participants (one survey entry per tank). Note that multiple uses were reported for each tank and percentages were calculated using the total number of study participants that responded to the question in each location (n = 76 and n = 44, for Currumbin and Brisbane, respectively). The survey was undertaken during March-September 2015...... 36
Table 3.2 Occurrence of fecal indicator bacteria (FIB) and potential opportunistic pathogens in Brisbane (n = 84) and Currumbin (n = 50) rainwater samples ...... 38
Table 3.3 Correlations between opportunistic pathogens and fecal indicators in roof-harvested rainwater. Significant (P < 0.05) values are bold-faced...... 43
Table 4.1 Characteristics and end uses of rainwater systems tested in this study...... 60
Table 4.2 Occurrence of fecal indicator bacteria and opportunistic pathogens (OPPPs) in tank water samples (n = 24) over six events ...... 64
Table 4.3 Correlations between fecal indicator bacteria (FIB) and opportunistic premise plumbing pathogens (OPPPs) in roof-harvested rainwater (RHRW) stored in tanks. Significant values (p < 0.031) are bold-faced; see methods for discussion of correcting for multiple comparisons using a FDR approach...... 69
Table 5.1 MAC primary host organisms and disease outcomes (adapted from Hibiya et al. (2011) and Rindi and Garzelli (2014) unless otherwise noted) ...... 83
Table 5.2. Data used in the current study for MAC dose response modeling ...... 96
Table 5.3. Model fit results for data in Table 5.2 (exponential model parameter r, Beta-
Poisson model parameters α, N50)...... 98
Table 6.1 Monte carlo simulation input parameters for ingestion scenarios (for MAC only) 145
Table 6.2 Monte carlo simulation input parameters for inhalation scenarios ...... 147
Table 6.3 Monte carlo simulation input parameters for pathogen concentrations ...... 149
Table 6.4 Monte carlo simulation dose response input parameters ...... 150 xiii
Table 6.5 Sensitivity analysis with Spearman rank correlation coefficients for ingestion risk scenarios. Risks for endpoints of children with cervical lymphadenitis and severe immune deficiency with disseminated infection are shown ...... 156
Table 6.6 Sensitivity analysis with Spearman rank correlation coefficients for inhalation risk scenarios. Risks for endpoints of pulmonary infection in healthy populations with L. pneumophila or MAC are shown...... 158
Table 7.1 Aerosol size distribution for modern flush toilets (Johnson et al. 2013) ...... 174
Table 7.2 Exposure parameters for toilet flushing scenarios ...... 175
Table 7.3 Pasquill Stability Classes for moderate solar radiation (Seinfeld 1986) ...... 180
Table 7.4 Concentrations distributions for Legionella in reclaimed and drinking water ...... 186
Table 7.5 Monte Carlo model input parameters ...... 191
Table 7.6 Risk characterization parameters ...... 193
Table 7.7 Comparison of annual risks across all sites for toilet flushing scenarios using reclaimed water (R) or treated drinking water (DW) for infection (Inf) or Death (D) animal dose response model endpoints using three different exposure models ...... 199
Table 7.8 Common sprinkler systems used for reclaimed water irrigation (NA=information not available from manufacturer) ...... 214
Table 9.1 Literature review search results for RHRW (Q=queried, I= imported to digital library) ...... 226
Table 9.2 Summary of design and maintenance for reviewed rainwater collection apparatuses ...... 228
Table 9.3 Rainwater harvesting practices from 136 survey participants in the United States adapted from Thomas et al. (2014) ...... 237
Table 9.4 Gene fragments used to construct plasmids for qPCR standards ...... 246
Table 9.5 qPCR primers, probes, and reaction mixtures used in quantitative PCR (qPCR) analysis ...... 247
Table 9.6 Rainwater tank characteristics reported by Brisbane (n = 76 total survey respondents) and Currumbin (n = 45 total survey respondents) rainwater survey participants (one survey entry per tank). Note that multiple uses were reported for each tank and percentages were calculated using the total number of study participants that responded to the question in each location (n = 76 and n = 44, for Brisbane and Currumbin, respectively) .... 248
Table 9.7 Mean ± standard deviation (SD), range of Amplification efficiencies (E), Correlation coefficient (r2) and Slope of the standard curves for qPCR assays...... 249
xiv
Table 9.8 Opportunistic premise plumbing pathogens (OPPPs) in 134 tank water samples in phase 1. Tank water samples chosen for phase 2 indicated (*) ...... 263
Table 9.9: qPCR primers, probes, and reaction mixtures used in quantitative PCR (qPCR) analysis ...... 268
Table 9.10 Mean ± standard deviation (SD), range of Amplification efficiencies (E), Correlation coefficient (r2) and Slope of the standard curves for qPCR assays...... 269
Table 9.11 List of meteorological stations closest to sampling sites 1-24 that provided rainfall only (Gauge A- L) or a complete set of meteorological data (Gauge 1-5) ...... 270
Table 9.12 Cochran Q test statistics and McNemar post-hoc testing results indicating which month(s) were significantly different. Significant results (p < 0.05 Cochran Q; McNemar post-hoc test p < 0.003, Bonferroni corrected value) shown in bold...... 271
Table 9.13 Friedman test statistics and Wilcoxon signed-rank post-hoc testing results indicating which month(s) were significantly different. Significant results (p < 0.05 Friedman test; Wilcoxon post-hoc test p < 0.003, Bonferroni corrected value) shown in bold...... 273
Table 9.14 Significant correlations between the presence or absence of FIB and pathogens with meteorological parameters, determined using binary logistic regression across all sampling events. Significant correlations (p < 0.016 and CI does not include 1) shown); see methods for discussion of correcting for multiple comparisons using a FDR approach. Rain Day 0 through -7 indicate days antecedent to the sampling event...... 275
Table 9.15 Correlations between microorganisms in roof-harvested rainwater with meteorological parameters. Significant values (p < 0.031) are bold-faced); see methods for discussion of correcting for multiple comparisons using a FDR approach. Rain Day 0 through -7 indicate days antecedent to the sampling event...... 277
Table 9.16 MAC quantification methods. A review of additional molecular (primarily non- quantitative) methodology is available in (Halstrom et al. 2015) and is recommended as a companion to this table...... 280
Table 9.17 Summary of systematic literature review for MAC dose response (Q=Queried, I= Imported) ...... 281
Table 9.18 In vivo dose response datasets for human-relevant MAC species ...... 282
Table 9.19 Environmental sources where MAC species were identified. For additional discussion of MAC isolated from animals, animal foods, animal and vegetable food products, and agricultural wastes, the reader is referred to (Pavlik et al. 2009c)...... 294
Table 9.20 Water uses reported by Currumbin (n = 45) and Brisbane (n = 76) rainwater survey participants (one survey entry per tank). Note that multiple uses were reported for each tank and percentages were calculated using the total number of study participants that responded to the question in each location (n = 76 and n = 44, for Currumbin and Brisbane, respectively).
xv
The survey was undertaken during March-September 2015 and reproduced from (Hamilton et al. 2016)...... 297
Table 9.21 Summary of Legionella exposure models reviewed ...... 303
Table 9.22 Risk characterization parameters for reviewed models that conducted full QMRAs for which a full model description was available ...... 326
xvi
List of Figures
Figure 2.1 Prioritization of pathogens in RHRW by exposure volume that would result in a 1× 10-4 annual infection risk or greater (per contact event) ...... 16
Figure 2.2 Sensitivity analysis for prioritization of pathogens in RHRW ...... 16
Figure 3.1 Box and whisker plots of the log concentrations of fecal indicators in positive samples from 134 rainwater tanks. The inner box lines represent the medians while the outer box lines represent the 25th and 75th data percentiles (Interquartile range, IQR), and the whiskers extend to the range. For each microorganism, dotted lines represent the lower limit of quantification of the respective assay...... 38
Figure 3.2 Box and whisker plots of the log gene copy concentrations of positive samples for each opportunistic pathogen in 134 rainwater tanks. The inner box lines represent the medians, while the outer box lines represent the 25th and 75th data percentiles (Interquartile range, IQR), and the whiskers extend to the range. For each microorganism, dotted lines represent the lower limit of quantification of the respective assay...... 41
Figure 4.1 Concentrations of fecal indicator bacteria (FIB) in roof-harvested rainwater (RHRW) from Queensland, Australia over six monthly sampling events. Red dotted line denotes the limit of quantification; horizontal hash marks denote the median concentration. Where no horizontal hash marks are present, the median was below the detection limit...... 65
Figure 4.2 Concentrations of opportunistic premise plumbing pathogens (OPPPs) in roof- harvested rainwater (RHRW) from Queensland, Australia over six monthly sampling events. Red dotted line denotes the limit of quantification; Green lines denote the limit of detection; horizontal hash marks denote the median concentration. Where no horizontal hash marks are present, the median was below the detection limit...... 66
Figure 4.3 Odds ratio and 95% confidence intervals for the ability of individual FIB and opportunistic premise plumbing pathogens (OPPPs) to predict the presence of other fecal indicator bacteria (FIB) and OPPPs in pooled tank water samples (n = 144). * indicates significance level (p < 0.017); see methods for discussion of correcting for multiple comparisons using a FDR approach. EC = E. coli, ENT = Enterococcus spp., ACAN = Acanthamoeba spp., LEG = Legionella spp., LP = L. pneumophila, MA = M. avium, MI = M. intracellulare, PA = P. aeruginosa...... 70
Figure 5.1 Mycobacterium avium complex (MAC) quantitative microbial risk assessment (QMRA) framework ...... 94
Figure 5.2. Exponential model fit for Tomioka 1993. Pooled M. avium lung lesions serovar 1 strains N-289, N-364, N-445, N-458, N-461 and serovar 9 strains N-254, N-302 ...... 99
Figure 5.3. Exponential model fit for Mehta 1996 for MAC 101- death at 30 days ...... 100
Figure 5.4. Exponential model fit for Jorgensen 1977 (Table 5.3 No. 1)- M. avium serotype 2 (pig origin)- 1. Lymph node lesions: mandibular, parotideus, retropharyng. lat., retropharyng. xvii
med., cervicalis sup. dors., cervicalis sup. ventr., subiliacus, popliteus, inguinalis prof., inguinalis superf., spleen. Lymph node Culture: cervicalis sup., cervicalis sup. ventr., subiliacus, ingninalis prof., igninalis superf.liver, kidney ...... 101
Figure 5.5. Exponential model fit for Jorgensen 1977 (Table 5.3 No. 2)- M. avium serotype 2 (pig origin) Lymph node Lesions: Liver, Intestinal mucosa (Peyer Patch) ...... 102
Figure 5.6. Exponential model fit for Jorgensen 1977 (Table 5.3 No. 3) M. avium serotype 2 Culture from retropharyng. lat. lymph nodes ...... 103
Figure 5.7. Yangco 1989 Beta poisson model fit to disseminated infection data ...... 104
Figure 5.8. Beta-Poisson model fit to Yangco 1989 data- liver infection ...... 105
Figure 6.1 Exposure routes for L. pneumophila and MAC in roof-harvested rainwater ...... 132
Figure 6.2 Annual cervical lymphadenitis (children) or disseminated infection (vulnerable/immune compromised populations) risks for ingestion of Mycobacterium avium complex through food or water exposure scenarios. Median and 95% confidence intervals and scenario analysis for various pool dilution levels (D = 10%, 50%, or 90% RHRW: sterile water), with or without drinking water under-sink point of use filtration (filtration/no filtration), and swimming frequency assumptions (32 or 122 swims per year) are shown .... 153
Figure 6.3 Annual infection risks for inhalation of Mycobacterium avium complex or L. pneumophila for various exposure scenarios. Median and 95% confidence intervals and scenario analysis for various pool dilution levels (D = 10%, 50%, or 90% RHRW: sterile water) and swimming assumptions (32 or 122 swims per year) are shown ...... 155
Figure 6.4 Total annual risks from all activities for cervical lymphadenitis in children or disseminated infection in vulnerable/immune compromised populations via ingestion. Median and 95% confidence intervals shown. Scenario analysis for various pool dilution levels (D = 10%, 50%, or 90% RHRW: sterile water), with or without drinking water under-sink point of use filtration (filtration/no filtration), and swimming frequency assumptions (32 or 122 swims per year) are shown...... 160
Figure 6.5 Total annual pulmonary infection risks from all activities via inhalation of Mycobacterium avium complex (MAC), L. pneumophila (LP), or both organisms (LP+MAC). Median and 95% confidence intervals are shown. Scenario analysis for various pool dilution levels (D = 10%, 50%, or 90% RHRW: sterile water) and swimming assumptions (32 or 122 swims per year) are shown...... 161
Figure 7.1 Annual risks for all and site-specific toilet flushing scenarios for infection (Inf) or Death (D) animal dose response model endpoints using three different exposure models. ... 197
Figure 7.2 Sensitivity analysis showing Spearman rank correlation coefficients for toilet exposure model methods 1, 2, and 3. Coefficients identify the most important predictive factors of annual infection or clinical severity infection risk, where 0 is no influence and -1 or +1 when the output is wholly dependent on that input...... 201
xviii
Figure 7.3 Annual risk (10y) of infection for Legionella pneumophila (culture method) in residential populations due to cooling tower exposure (stack height= 10 m, efficiency 0.001- 0.005%) at varying downwind distances and meteorological parameters. The median (solid black line) and 5th, 95th percentiles (dotted lines) are shown...... 204
Figure 7.4 Annual infection risks (10y) for Legionella pneumophila in residential populations due to cooling tower aerosol exposure (efficiency = 0.001-0.005%) at varying downwind distances for wind speed = 7 m / s and relative humidity = 65%. The median (solid black line) and 95% confidence interval (dotted lines) are shown ...... 206
Figure 7.5 Annual infection risks (10y) for Legionella pneumophila in occupational populations due to cooling tower aerosol exposure at varying downwind distances for wind speed = 7 m / s, relative humidity = 65%, and stack height=10 m. The median (solid black line) and 95% confidence interval (dotted lines) are shown ...... 207
Figure 7.6 Sensitivity analysis for cooling towers ...... 209
Figure 7.7 Annual risk (10y) of infection for Legionella pneumophila (culture method) in residential populations due to sprinkler aerosol exposure at varying downwind distances and meteorological parameters. The median (solid black line) and 5th, 95th percentiles (dotted lines) are shown ...... 211
Figure 7.8 Annual infection risks (10y) for Legionella pneumophila in residential populations due to sprinkler exposure at varying downwind distances for wind speed = 7 m / s and relative humidity = 65%. The median (solid black line) and 95% confidence interval (dotted lines) are shown...... 212
Figure 7.9 Annual infection risks (10y) for Legionella pneumophila in occupational populations due to sprinkler aerosol exposure at varying downwind distances for wind speed = 7 m / s, and relative humidity = 65%. The median (solid black line) and 5th, 95th percentiles (dotted lines) are shown...... 213
Figure 7.10 Sensitivity analysis for sprinklers ...... 216
Figure 9.1 Recruitment letter ...... 250
Figure 9.2 Voluntary survey sent to rainwater study participants ...... 256
Figure 9.3 Email text used to disseminate results to rainwater study participants ...... 259
Figure 9.4 Template for personalized results sheet provided to each rainwater study participant in screening and seasonal study ...... 262
Figure 9.5 Total monthly rainfall data for gauges closest to Brisbane (Gauge G) and the Gold Coast (Gauge L) ...... 278
Figure 9.6 Daily rainfall over six month sampling period for two gauges representative of Brisbane (Gauge G) and Gold Coast (Gauge L) sites...... 278
xix
Figure 9.7 Daily minimum and maximum temperatures observed at five gauges closest to the sampling sites where full meteorological datasets were available ...... 279
Figure 9.8 Daily relative humidity measurements taken at 9am and 3pm at five gauges closest to the sampling sites where full meteorological datasets were available...... 279
Figure 9.9 Proposed comprehensive framework for Legionella quantitative microbial risk assessment ...... 337
Figure 9.10 Annual infection risks (10y) for Legionella pneumophila in residential populations due to cooling tower aerosol exposure using drinking water at varying downwind distances for wind speed = 7 m / s and relative humidity = 65%. The median (solid black line) and 95% confidence interval (dotted lines) are shown...... 338
Figure 9.11 Annual infection risks (10y) for Legionella pneumophila in occupational populations due to cooling tower aerosol exposure using drinking water at varying downwind distances for wind speed = 7 m / s and relative humidity = 65%. The median (solid black line) and 95% confidence interval (dotted lines) are shown...... 339
Figure 9.12 Annual infection risks (10y) for Legionella pneumophila due to cooling tower aerosol (0.001- 0.005% efficiency) exposure at varying downwind distances for wind speed = 7 m / s, relative humidity = 65%, stack height=10 m, infection dose response endpoint. The median (solid black line) and 95% confidence interval (dotted lines) are shown. Note that in some cases the 5th percentile did not converge due to a low fraction of samples positive..... 341
Figure 9.13 Impact of cooling tower drift efficiency (left: 0.001-0.005%; right: 0.01-0.1%) on annual infection risks (10y) for Legionella pneumophila in reclaimed water for residential populations due to cooling tower aerosol exposure at varying downwind distances for culture data, wind speed = 7 m / s, relative humidity = 65%, and stack height=10 m. The median (solid black line) and 95% confidence interval (dotted lines) are shown...... 342
Figure 9.14 Annual infection risks (10y) for Legionella pneumophila in residential populations due to sprinkler aerosol exposure using drinking water at varying downwind distances for wind speed = 7 m / s and relative humidity = 65%. The median (solid black line) and 95% confidence interval (dotted lines) are shown...... 343
Figure 9.15 Annual infection risks (10y) for Legionella pneumophila in occupational populations due to sprinkler aerosol exposure using drinking water at varying downwind distances for wind speed = 7 m / s and relative humidity = 65%. The median (solid black line) and 5th, 95th percentiles (dotted lines) are shown...... 344
Figure 9.16 Annual infection risks (10y) for Legionella pneumophila due to sprinkler aerosol exposure at varying downwind distances for wind speed = 7 m / s, relative humidity = 65%, infection dose response endpoint. The median (solid black line) and 5th, 95th percentiles (dotted lines) are shown. aNote that in some cases the 5th percentile did not converge due to a low fraction of samples positive...... 346
xx
Abstract
Quantitative microbial risk assessment for sustainable water resources
Kerry Ann Hamilton, M.H.S, C.P.H., E.I.T.
Charles N. Haas, Ph.D.
This thesis evaluates the microbiological health risks of using the sustainable water resources roof-harvested rainwater and reclaimed water. This research was accomplished by performing six tasks: 1) prioritizing pathogens for exploration; 2) conducting a screening study of opportunistic pathogens in rainwater tanks identified in task 1 in Brisbane, Australia; 3) conducting a longitudinal study of opportunistic pathogens for contaminated tanks identified in task 2; 4) developing a series of new dose response models necessary to conduct a risk assessment; 5) conducting a quantitative microbial risk assessment (QMRA) for two index opportunistic pathogens in roof-harvested rainwater; and 6) conducting a QMRA for an opportunistic pathogen in reclaimed water in partnership with a large field and laboratory study conducted by American Water.
Decentralized, alternative, or otherwise “sustainable” water resources are growing in popularity as the world’s water resources are strained due to population growth, climate change, and water scarcity. These resources bring new challenges for the water industry in terms of maintaining water quality standards and minimizing adverse impacts. Increased attention has been devoted to opportunistic pathogens due to their growing importance as a portion of the waterborne disease burden in many countries and need to assess their associated health risks. The two systems xxi investigated in this thesis are roof-harvested rainwater (RHRW) systems and reclaimed water. Roof-harvested rainwater systems are used in many parts of the world to collect rainwater that falls on roof surfaces in a tank or barrel. This rainwater is then used for a variety of potable or non-potable purposes. Reclaimed water is wastewater reused for beneficial purposes with treatment, where the level of treatment depends on the reuse application.
Both types of water systems can foster environments that are conducive to the occurrence of opportunistic pathogens. Opportunistic pathogens are microorganisms that are pathogenic under certain sets of circumstances and typically affect children, the elderly, and immune-compromised hosts rather than healthy individuals. The two opportunistic pathogens that are the focus of this work are Legionella pneumophila and Mycobacterium avium complex (MAC). L. pneumophila causes infection and respiratory disease through inhalation of aqueous aerosols, while MAC can causes infection or disease through inhalation or ingestion. Due to the use of RHRW for both potable and non-potable purposes, and the use of reclaimed water for potable purposes that generate large-scale aerosols (such as cooling tower mists), the potential for public health risk exists as a result of human contact with these water sources. A
QMRA is therefore needed to assess risk and prioritize risk management and data gathering needs for both of these pathogens under a variety of scenarios.
This thesis presents a large-scale field study of RHRW in Brisbane, Australia, where
RHRW is used on a large scale due to aggressive droughts in the region. Molecular biology methods (qPCR) and culture-based methods were used to screen tanks for opportunistic pathogens, and a subset of contaminated tanks was chosen for a six-
xxii month follow-up study. A survey of rainwater tank owner use and maintenance practice was conducted. Additionally, an analysis of the correlation among microorganisms and between microorganisms and meteorological factors was performed to inform risk management approaches. A major barrier to conducting a
QMRA for MAC has been the development of a risk assessment framework and dose response models. These models have been developed for this work and as a result, a
QMRA is performed for RHRW using Monte Carlo simulation and sensitivity analysis. A similar modeling approach is used for assessing risks from toilet flushing, spray irrigation, and cooling tower use for reclaimed water. As a result, appropriate uses for RHRW are designated and appropriate setback distances for cooling towers and spray irrigation systems are proposed for reclaimed water.
1
1. Introduction
1.1. Background
As new methods for conserving water and mitigating climate change in urban areas have become more prevalent, the potential for increased public contact with water of variable quality is of potential concern. Although the hydrologic benefits of roof-harvested rainwater and reclaimed water have been studied, the health risks from exposure to microbial pathogens in these systems have not been systematically analyzed and compared to determine relative circumstances that minimize risk for each scenario. The goal of this work is to quantify these risks in order to encourage sustainable water use while protecting public health.
A literature review was performed and is presented here for contaminant occurrence from roof-harvested rainwater (RHRW) and reclaimed water, as well as information on human exposure and disease. The literature review identifies
Legionella pneumophila and Mycobacterium avium complex (MAC) as two pathogens of priority public health focus. These pathogens can also amplify within protozoans and modulate human virulence within dynamic microbial ecosystems found within water pipes, storage tanks, and natural ecosystems. This work formed the basis for field studies of Legionella and MAC in RHRW and reclaimed water.
This work develops new dose response models for MAC, enabling the use of quantitative microbial risk assessment (QMRA) as a tool to characterize and compare health risks from these systems.
All RHRW-related work for this thesis was performed in Brisbane, Australia from
January 2015- April 2016 and the author performed all related laboratory analysis.
Risk models for RHRW were developed for Australian populations. For reclaimed 2 water, laboratory analysis was performed by American Water for United States water utilities. All risk models for reclaimed water are focused on United States populations.
3
1.2. Research Objectives
The purpose of this thesis is to quantify health risks from opportunistic pathogens in two types of water systems designed to increase water conservation and sustainability. This is accomplished by (1) conducting field studies and laboratory analysis to quantify pathogens; and (2) conducing Monte Carlo simulations to calculate microbiological health risks. Specific objectives and the thesis structure are discussed below.
1.2.1. Objective 1: Prioritize pathogens previously quantified in
harvested rainwater for further study
A variety of microbiological and chemical contaminants have been observed in harvested rainwater tanks (Abbasi and Abbasi 2011). Microbiological contaminants are the focus of this work due to the potential for acute health effects. Given the potential for health risks from multiple microorganisms, it is useful to systematically prioritize them according to established criteria for selecting reference hazards for further study. To achieve this goal, a systematic literature review was performed for concentrations of microbiological contaminants in RHRW in order to build a database of concentrations. A Monte Carlo simulation was then performed to assess the exposure volume necessary to incur an
“unacceptable” risk. The exposure volumes computed for each pathogen were used to rank them according to their relative potential for harm so that representative pathogens could be selected for laboratory and field study. Objective 1 is discussed in Chapter 2. Supplemental material for this section is provided in Chapter 9 section 9.1.
1.2.2. Objective 2: Conduct a screening of Brisbane rainwater tanks for
pathogens
The pathogens chosen from the systematic literature review in the previous section belong to the class of “opportunistic pathogens” which primarily causes disease in immune- compromised hosts, but are known to occur in high concentrations in engineered water systems under certain conditions. As a result, seven opportunistic pathogen targets
(Legionella spp., L. pneumophila, L. longbeachae, Acanthamoeba spp., M. avium, M. 4
intracellulare, P. aeruginosa) and two fecal indicator bacteria (E. coli, Enterococcus spp.) were quantified in a large field study of 134 RHRW tanks from Brisbane, Australia.
Participants were surveyed for RHRW maintenance, treatment, and use practices.
Correlations were assessed among microbiological targets. Objective 2 is discussed in
Chapter 3 and is published in Environmental Research. Supplemental material is provided in
Chapter 9 section 9.2.
1.2.3. Objective 3: Conduct a seasonal study of Brisbane rainwater tanks
for pathogens
To assess the variability of pathogen occurrence in RHRW tanks over time, a subset of 24 tanks in which FIB or opportunistic pathogens were present was selected for six-month follow-up study in Brisbane, Australia. In addition to tank sampling and pathogen quantification, meteorological data was obtained from the Australian Bureau of Meteorology for correlation with pathogen occurrence data. Objective 3 is discussed in Chapter 4 and published in Environmental Science & Technology. Supplemental material is provided in
Chapter 9 section 9.3.
1.2.4. Objective 4: Develop a dose response model for the Mycobacterium
avium complex (MAC)
Mycobacterium avium complex (MAC) was quantified in a large portion of RHRW tanks in Brisbane, Australia, warranting the assessment of health risks associated with this group of opportunistic pathogens. In order to complete a quantitative microbial risk assessment, it is necessary to have a dose response model for the microorganism. However, such a model was not available for the most human-relevant MAC species. As a result, a systematic literature review for experimental in vivo animal models was conducted in order to develop dose response models. This analysis revealed that risk analysis for MAC must take into account exposure routes and specific susceptible populations. A QMRA framework and seven new dose response models are proposed. Objective 4 is discussed in Chapter 5 and is published in Water Research. Supplemental material is provided in Chapter 9 Section 9.5.
5
1.2.5. Objective 5: Assess the health risks of Legionella and MAC in
harvested rainwater
While drinking water guidelines clearly specify that Australian rainwater should contain less than one fecal indicator bacteria per 100 mL of water, there is no consensus about which end-uses of rainwater are appropriate with regards to opportunistic pathogen health risks. Using the newly developed dose response models for MAC, a QMRA is performed to assess health risks from index pathogens L. pneumophila and MAC for a variety of RHRW inhalation and ingestion exposure scenarios. For inhalation, showering, pool top-up, gardening (hosing), car washing, and toilet flushing were considered. For ingestion, drinking, raw produce consumption, showering, gardening (hosing), pool top-up, car washing, toilet flushing, and clothes washing were considered. Performing this full suite of QMRA analyses allows for determination of appropriate uses. Objective 5 is discussed in Chapter 6 and supplemental material is provided in Chapter 9 section 9.4. This material has been submitted to a journal for review.
1.2.6. Objective 6: Assess the health risks of Legionella in reclaimed
water
A survey of 19 reclaimed water utilities from across the United States conducted by
American Water determined the presence of pathogenic species belonging to the genus
Legionella. In order to inform appropriate risk management practices, a QMRA was undertaken for three uses of reclaimed water: toilet flushing, spray irrigation, and cooling tower recirculating water systems that produce mist plumes. For spray irrigation and cooling towers, the purpose of this risk assessment was to predict an appropriate setback distance from the mist-generating device in order to minimize public health risks. Objective 6 is discussed in Chapter 7. A review paper conducted as preliminary analysis for this work was also published in Environmental Science: Water Research & Technology and is provided in
Chapter 9 section 9.6. Portions of Chapter 7 are published as WateReuse Research
Foundation Report WRF-05 “Development of a Risk Management Strategy for Legionella in
6
Recycled Water Systems”, in press. Supplemental material for Chapter 7 is provided in
Chapter 9 section 9.7.
7
2. Prioritization of pathogens in harvested rainwater
2.1. Roof-harvested rainwater
2.1.1. Introduction
Epidemiologic studies indicate that roof-harvested rainwater (RHRW)
(re)used for drinking or domestic use has been associated with disease risks and several outbreaks (Ahmed et al. 2011a, Ahmed et al. 2010, Brown et al. 2001, Dean and Hunter 2012, Eberhart-Phillips et al. 1997, Fewtrell and Kay 2007b, Heyworth et al. 2006, Lye 2002, Simmons et al. 2008). In urban areas of the United States, rainwater harvesting is currently being practiced and is increasing in some areas as a result of a growing awareness of water issues as well as incentives by various organizations (Mandarano 2011, NYCDEP 2011). However, RHRW systems are not federally regulated, and state or local regulations can vary substantially (USEPA
2013). Given the variable nature of RHRW quality from different roof surfaces, it is challenging to designate appropriate uses. Despite these limitations, it is useful to identify common contaminants and contributing factors for guiding rainwater treatment and (re)use.
A typical engineered rainwater collection system relies upon a sloped catchment system (e.g., a roof) which drains via a gutter and attached pipe into a storage container called a cistern (Farreny et al. 2011). The first flush is typically collected separately while overflow is diverted into a sewer system (Jordan 2008).
The remaining collected water can be pumped either into a treatment system supplying water for domestic use, or to an area where it can be used directly for irrigation or other use (Jordan 2008). The factors determining the volume of runoff 8 from the roof and storage of harvested water are similar to ground surface runoff and include the size of the roof, the size of the collecting area, the depth of rainfall, the runoff efficiency, the volume of runoff, and the volume of storage (Boers and Ben-
Asher 1982). The quality of this water varies according to geographic location, catchment location, climatic conditions, organic material in the guttering, the volume and retention time of the water in the tank, the roof condition, condition of the piping and storage systems, maintenance and management of the system (Thurman 1995,
Vialle et al. 2011).
In light of potential waterborne exposures, a number of studies have characterized the quality of RHRW, specifically in identifying suitable uses. RHRW contains various contaminants including microorganisms, chemicals, nutrients, and heavy metals. Comprehensive reviews of microbial rainwater quality from various locations are provided by Abbasi and Abbasi (2011) (global) Struck (2011) (UK),
Chapman (2008) (Australia), Meera and Ahammed (2006) (global), Ahmed et al.
(2011a) (global), Lye (2009) (global), (Fewtrell and Kay 2007a) (developed countries) and de Kwaadsteniet et al. (2013). In RHRW from North America and
Europe, the most common organisms are E. coli, Enterococci, total coliforms,
Cryptosporidium, Giardia, Pseudomonas aeruginosa, and Clostridium perfringens
(Abbasi and Abbasi 2011). A systematic literature review was performed in order to identify which of these pathogens represents the greatest health risk and should be studied further.
2.1.2. Human health Risks
C. botulinum, Campylobacter spp., S. typhimurium, L. pneumophila,
Cryptosporidium, and Giardia have previously been implicated in outbreaks
9 associated with RHRW (Ahmed et al. 2011a, Ahmed et al. 2010, Brown et al. 2001,
Dean and Hunter 2012, Eberhart-Phillips et al. 1997, Fewtrell and Kay 2007b,
Heyworth et al. 2006, Lye 2002, Simmons et al. 2008). In addition, several studies have used water quality information to quantify the microbial health risks associated with the (re)use of water from rainwater catchment systems including for consumption of untreated water via drinking or hosing (Ahmed et al. 2010, Ahmed
2009, Lye 2002) and aerosol ingestion or inhalation via hosing, showering, or flushing toilets with untreated RHRW (Ahmed et al. 2010, Ahmed 2009, Fewtrell and
Kay 2007b). A Southeast Queensland study reported the largest infection risks were presented by Salmonella spp. (9.8 – 53 cases per 10,000 persons/yr) and G. lamblia
(20- 130 cases / 10,000 persons/yr) (Ahmed et al. 2010). Other exposures for L. pneumophila, Salmonella spp. and G. lamblia were examined including liquid ingestion via hosing, aerosol ingestion via showering, and aerosol ingestion via hosing but demonstrated lower risks ranging from .01 to .24 cases /10,000 persons/ year (Ahmed et al. 2010). Fewtrell and Kay (2007b) examined the microbial risks associated with Campylobacter infection from toilet flushing with rainwater in the
UK and estimated a mean probability of infection of 1.8 × 10-5 (range 3.7 × 10-9 –
1.1× 10-4), mean probability of illness of 5.4 × 10-6 (1.1×10-9- 3.4×10-5), and mean cases of campylobacteriosis of 0.023 (.14 - 4.64 × 10-6). These probabilities were quantified in terms of disability adjusted life years (DALYs) of 6.8 ×10-5 (1.4× 10-8 -
4.0 × 10-4), approximately 2 orders of magnitude below the World Health
Organization (WHO) standard (Fewtrell and Kay 2007b). No risk assessment was located to date quantifying risks from exposure to chemical or metal contaminants in
RHRW. However, drinking cistern water has been associated with a lower incidence
10 of colon and rectal cancers in Northern Kentucky compared to drinking water from municipal water systems in an early study (Lye 1992, Richmond et al. 1987).
2.1.2.1. Methodology
A systematic literature review was conducted for concentrations of microorganisms, chemicals, nutrients, and metals found in RHRW using combinations of keywords (Appendix Table 9.1) in Google Scholar, Proquest
Environmental Engineering Abstracts, Agricola, Engineering Village, and Web of
Knowledge databases. Inclusion criteria were 1) the paper originates from a developed country as classified by the US Central Intelligence Agency World
Factbook (https://www.cia.gov/library/publications/the-world- factbook/appendix/appendix-b.html ); 2) peer-reviewed; 3) in English; 4) contains information on the presence, absence, or quantitative concentration of microorganisms. Forward and reverse citation searches were used for records containing metals, chemicals, or microorganism concentration values for RHRW or direct rainfall. Previous literature reviews were identified, re-reviewed, and forward/reverse citation searched for factors influencing RHRW quality. All original papers referenced in reviews were consulted including runoff from traditional, green, or cool roofs. Abstracts were reviewed and full text for relevant PDFs meeting selection criteria were obtained, formatted with Adobe Acrobat X Pro ® for text recognition, and imported into an EndNote® library. EndNote Smart Groups were used to further focus studies using relevant keywords. Sixteen studies not previously reviewed are summarized here. Concentrations of contaminants were excerpted and compiled in an Excel spreadsheet. Where data were not available in tabular form, it was extracted from images using Digitize It® 2010 Version 4.0.2.
11
A health risk assessment framework was used to prioritize contaminants based on the concentrations summarized in this review. Only the analysis from microorganisms is presented here. Dose response information for microorganisms was obtained from the Center for Advancing Microbial Risk Assessment (CAMRA) wiki
(http://qmrawiki.msu.edu).
Microbial contaminants were quantitatively prioritized based on the exposure volume necessary to generate an infection risk of 1 in 10,000 (the USEPA recreational water criteria) by solving exponential (equation 2.1) and beta-Poisson (equation 2.3) dose-response models (Haas et al., 1999) for the dose necessary to generate this risk
(equation 2.2,2.4).
푃(푑) = 1 − 푒−푟푑 Equation 2.1
ln(1−푃(푑)) 푑 = Equation 2.2 −푟
푑 푃(푑) = 1 − (1 + )−훼 Equation 2.3 훽
1 1 − 푑 = 훽[( ) 훼 − 1] Equation 2.4 1−푃(푑)
Where P(d) is the probability of response at dose “d”, “r” is the probability that a single organism can survive and initiate infection, and α, β are parameters of the beta-
Poisson model.
A Monte Carlo risk model was developed using @Risk (©Palisade, 2012) software. Input distributions were developed from literature values using best-fit dose
12 response parameters for human oral/inhalation infection (qmrawiki.msu.edu) and pathogen concentrations. The dose distribution for each pathogen was multiplied by the reciprocal of its respective concentration distribution to obtain a volume.
Maximum likelihood estimation (MLE) was used to determine the parameters of a pooled distribution of concentrations (mean of maximum of reported ranges) where appropriate as in Pepper et al. (2010) using weighting based on the number of samples and censoring at an approximate limit of detection (1/sample volume). Studies reporting only percent positive samples were converted to MPN values using standard formulae in Haas et al. (1999). Dose response parameters used to prioritize the organisms are listed in Table 2.1. Dose response models were not available for
Aeromonas, Clostridium perfringens, Mycobacterium avium Complex (MAC),
Streptococcus, or Pseudomonas syringae.
13
Table 2.1 Best-fit dose response model parameters for pathogenic microorganisms reported in RHRW (qmrawiki.canr.msu.edu)
Organism Host Strain Best fit parameters
Legionella pneumophila Guinea pig Philadelphia 1 k=5.99×10-2 (4.18 ×10-2, 1.57×10-1) Cryptosporidium parvum Human TAMU isolate k=5.72×10-2 (2.46×10-2, 2.65) Giardia duodenalis Human Human-derived k=1.99×10-2 (1.26×10-2, 2.92×10-2) Mycobacterium avium Deer Paratuberculosis- k=6.93×10-4 (3.75×10-4, 1.16×10-3) Bovine E. coli (EPEC) Human E2348/69 (O127:H6) k=1.95×10-6 (1.25×10-6, 2.64×10-6) E. coli (ETEC) Human B7A α=3.75×10-1 (1.34×10-1, 99.7), N =1.78×105 (3.63×10-1, 50 2.46×106) Salmonella nontyphoid Human Multiple (Haas et al. α=0.3126, N =2.36×104 1999) 50 Campylobacter jejuni Human A3249 α= 1.44×10-1 (5.15×10-2, 2.68×10-1) N = 8.9×102 (1.87×10-1, 7.15×103) 50 14
2.1.2.2. Results/ literature review findings
A comprehensive review of factors that affect RHRW quality is provided in
Appendix section 9.1. Pathogens present in sixteen North American/European studies that had not been previously reviewed are summarized and compared to estimates from previously reviewed studies in Table 2.2. Characteristics of the reviewed studies and rainwater use practices are summarized in Appendix Table 9.2- Table 9.3. 15
Table 2.2 Ranges of microbial contaminants reported for RHRW across multiple studies (V= vegetated roofs, NV= non-vegetated roofs). Original references are provided in Chapter 9 Appendix Table 9.2.
Microorganism Ranges in previously reviewed studies Ranges in 16 new studies (#/100mL) NV V NV V 10 studies No studies 9 studies 2 studies Aeromonas spp. Max: 300 No studies Mean: 31-15,642 Mean: 1,162- Range: 0- 85,000 4,193 Range: 70- 16,818 C. perfringens No studies No studies Mean: 0-13 Mean: 1-3 Range: 0-43 Range: 0-5 Campylobacter spp.
95% confidence intervals for the volumes necessary to incur a 1e-4 infection risk are presented in Figure 2.1. C. parvum, G. duodenalis, L. pneumophila, and M. avium exceeded the target risk for volumes less than 1 L based on reported concentrations of
RHRW pathogens. These data indicate that at low volumes, the potential for exceeding a drinking water target risk of 1e-4 exists and provides motivation for 16 further study of occurrence and exposure patterns of RHRW. Concentration uncertainty dominated dose response uncertainty in 3/5 modelled scenarios (Figure
2.2).
1.00E+05 1.00E+04 1.00E+03
1.00E+02
4 4 Probability - 1.00E+01 1.00E+00 1.00E-01
of Infection of 1.00E-02 1.00E-03 1.00E-04
1.00E-05 Exposure volume for (L) 1E 1.00E-06 C. parvum G. duodenalis L. pneumophila M. avium Salmonella spp. C. jejuni Pathogen Figure 2.1 Prioritization of pathogens in RHRW by exposure volume that would result in a 1× 10-4 annual infection risk or greater (per contact event)
C. parvum
L. pneumophila r
G. duodenalis Concentration N50 M. avium Alpha
C. jejuni
-1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6
Figure 2.2 Sensitivity analysis for prioritization of pathogens in RHRW
17
2.2. Pathogens identified for further study
2.2.1. Legionella pneumophila
Legionella is a waterborne pathogen of significant public health importance known to occur in engineered water systems, ambient water environments and soils. It is atypical in that it grows in biofilms in piping and can slough off and become aerosolized through water fixtures. Legionnaires’ disease (LD) or legionellosis results from inhalation of aqueous aerosols containing Legionella bacteria. The most common cause of illness is Legionella pneumophila, although over 50 species have been identified (Diederen 2008). Infections are particularly problematic for vulnerable and/or susceptible populations such as those who are immunocompromised/ immunosuppressed, the elderly, smokers, and other hospitalized individuals
(Fliermans 1996, Marston et al. 1994). Although rapid diagnosis and treatment methods have improved case-fatality rates since the first recognized outbreak in 1976, prevention of LD remains an important focus for water quality management (Benin et al. 2002). A summary of literature relevant to Legionella and climate change is provided here, as well as a brief discussion of global outbreaks and recommendations for future research.
2.2.1.1. Survival
Survival of Legionella in liquid aerosol is impacted by temperature, humidity, growth/metabolic activity conditions and composition of the aerosol. Bartram et al.
(2007) provides a detailed discussion of the impact of temperature on Legionella pneumophila, which generally multiplies in the 20°C-50°C range. An early study by
Berendt (1980) reported increased stability of Legionella pneumophila Philadelphia-1
18 strain survival in aerosols of tryptose-saline solution at 80% relative humidity (RH) compared to 50% or 30%- RH. Hambleton et al. (1983) reported similar findings except that organisms survived best at 65% RH for L. pneumophila strain 74/81 in distilled water. Aerosolized Legionella are more stable (with no statistical effect on viability) in solution with cyanobacterial extracts or nutrients, suggesting stabilization of Legionella under these conditions in natural environments (Berendt 1981, Tesh and
Miller 1981, Warren and Miller 1979).
2.2.1.2. Ecology
Legionella displays a complex microbial ecology and can reside freely or within a protozoan host (Lau and Ashbolt 2009a). It can replicate in amebae at temperatures
>25°C and digest amoeba at <20°C, highlighting the importance of temperature as a factor for growth directly as well as indirectly via impacts on biofilm formation
(Ohno et al. 2008). A comparison of natural water samples upstream and downstream from thermal baths and a wastewater treatment plant from Tech River, France showed
Legionella species diversity changed in relation to the thermal bath, but only seasonally for the wastewater treatment plant (Parthuisot et al. 2010). Although only
L. pneumophila could be cultured, qPCR assays showed large concentrations (up to
9.36 x 105 genome units (GU) * L-1, approaching concentrations seen in cooling towers) of multiple Legionella species in natural waters (Parthuisot et al. 2010). More research on the ecology of Legionella in the natural water environment is needed to understand their sources and entry into engineered water systems (Parthuisot et al.
2010). Legionella has also been isolated in rain water puddles (7/18 samples), particularly at temperatures greater than 25°C, suggesting additional exposure
19 scenarios due to potentially entering a replicative state at these higher temperatures
(Sakamoto et al. 2009). Differences in engineered system occurrence are noted between facilities with year-round (0/11 positive facilities) versus seasonal (4/10 positive facilities) habitation, with no significant difference in chlorine concentration or metals other than higher zinc concentrations observed in year-round accommodations (Rakić et al. 2011).
2.2.1.3. Global outbreaks.
Reporting for Legionella began in the United States in 2001. From 2001- 2006,
Legionella was identified as the causative agent of 38 of 833 U.S. drinking water outbreaks from all water types (389 cases) and from 1971- 2006 was identified as the cause of all reported drinking water-associated acute respiratory illness (ARI) (Craun et al. 2010). In 2005-2006, ARI surpassed acute gastrointestinal illness (AGI) as the leading cause of waterborne illness from drinking water in the U.S. (Yoder et al.
2008a). During this period, eight recreational water outbreaks (all from whirlpool spas) were linked to Legionella, with 124 cases and three deaths (Yoder et al. 2008b).
In 2007 and 2008, Legionella was the most frequently reported cause of drinking water-associated outbreaks in the U.S. (Brunkard et al. 2011, Craun et al. 2010). Ten spa-associated outbreaks were identified during this period (122 cases) (Hlavsa and
Brunkard 2011). From 1990-2005 this included a total of 23,076 cases of legionellosis, with most cases reported in summer or fall (Neil and Berkelman 2008).
Major European LD outbreaks and clusters up to 2004 and 2005 are reported in
Bartram et al. (2007) and Joseph and Ricketts (2008), respectively. From 1995 to
20
2005, 656 outbreaks (over 32,000 cases) of legionellosis from 35 European countries were reported to The European Working Group for Legionella Infections (EWGLI)
(Joseph and Ricketts 2008). These outbreaks were associated with nosocomial infection (100), community acquired pneumonia (160), travel (390) and other causes
(6), with the most frequently identified major sources reported as water systems (207), cooling towers (57), and spa pools (33) (Joseph and Ricketts 2008). From 2005-2006,
214 outbreaks were reported (11,980 cases) across 35 countries with a case fatality rate (CFR) of 6.6% (Ricketts 2007). Joseph and Ricketts (2010) report 5,907 cases of
LD across 33 European countries in 2007 and 5,960 cases across 34 countries in 2008 with the same two-year CFR. Additional outbreak data published since 2006 (EWGLI outbreaks since 2008) are reported in Table 2.3. Although diagnostic and analytical capabilities may vary globally, Legionella has been isolated in many other non- western countries including South Africa (Grabow et al. 1991), China (Lin et al.
2009), Namibia (Chinsembu and Hakwenye 2010), Nigeria (Alli et al. 2011), Japan
(Amemura-Maekawa et al. 2010), Thailand (Travis et al. 2012), Singapore (Lim et al.
2011), and Israel (Yarom et al. 2010).
21
Table 2.3 Outbreaks/clusters of LD 2006-present (NA= Not Available)
Source Location Outbreak Start Cases (Deaths) Attack Rate (%) Reference Air scrubber Norway 2005 56 (10) Sarpsborg-.027; Fredrikstad-.064; by (Nygård et al. 2008) source radii up to .272 b Bathhouse Japan 2002 295, 20 confirmed (4)a Confirmed cases-.101; Potential (Sasaki et al. 2008) cases- 1.49c Cooling system (suspected) Germany 2009 65(5) NA (Von Baum et al. 2010) Cooling tower (suspected) U.K. 2012 50(2), 49 additional suspected NA (McCormick et al. 2011) (1) Cooling tower Spain 2006 146 (0) Pamplona-.063; By section: Pamplona (Castilla et al. 2008a) up to .256; District 2 up to 1.23 Cooling tower Spain 2005 55 (3) Overall- .088; Vic-.087; Gurb-.091; (Ferre et al. 2009) N. Vic (within 2,000 m of suspected source)-.113; S. Vic- .042 Cooling tower Sweden 2004 30 (2) .12c (Hugosson et al. 2007) Cooling tower France 2003 86 (18) .039 (overall)-.167 (Harnes)b (Nguyen et al. 2006) Cooling tower Ireland 2008 2 (NA) NA (Nicolay et al. 2010) Cooling tower Spain 2002 113 (2) .400 (total); up to .510 by radiib (Sabria et al. 2006) Decorative fountain- hospital U.S. 2010 8 (0) 0.2c (Haupt et al. 2012) Decorative fountain- hospital U.S. 2007 2 (0) NA (Palmore et al. 2009) Decorative fountain- restaurant U.S. 2005 18 (1) NA (O'Loughlin et al. 2007) Distribution system- hotel Spain 2011 18(4) NA (Vanaclocha et al. 2012) Distribution system- nursing home Slovenia 2010 15, 4 confirmed (0) 6.41 (Skaza et al. 2010) Gardening/ potting soil Scotland 2008 3(1) NA (Pravinkumar 2010) Long term care facility Canada 2005 82 (23) NA (Gilmour et al. 2007) Military recruit facility U.S. 2004 5 (0) NA (McDonough et al. 2007) Nosocomial Iran 2007 6 (NA) .333c (Doust et al. 2009) Nosocomial Poland 2006 4 (3) 15 (Stypułkowska-Misiurewicz et al. 2006) Nosocomial Spain 2005 12 (1-attributable cases) 4c (Gudiol et al. 2007) Nosocomial Turkey 2004 7 (NA) 1.4c Ozerol et al. 2005 Rainwater system, marina water blaster New Zealand 2006 4 (1) NA (Simmons et al. 2008) Storage tank U.K. 2004 7 (1) NA (Hyland et al. 2008) Supermarket mist machine Spain 2006 12 (0) Vilafranca- .014 (Barrabeig et al. 2010) Supermarket- .020 Fish section- .074 Spa on cruise ship- porous stones in spa filter Japan 2003 3 (0) 1.5c (Kura et al. 2006) Spa pool U.K. 2012 21(NA) NA (Coetzee et al. 2012) Whirlpool spa/ hair wash station on cruise ship Iceland 2003 8 (1) 4 (Beyrer et al. 2007) (a)Not all suspected cases were examined; 295 suspected cases, 162 agreed to participate in epidemiological study, 76 met epidemiological and clinical criteria for probable cases, 20 confirmed by positive response to at least one Legionella-specific diagnostic test (b) Attack rates for multiple areas reported (c) Calculated here based on potentially exposed population
22
2.2.1.4. Seasonality and meteorological factors.
Despite increases in LD incidence, it continues to be difficult for public health practitioners to interpret surveillance data with respect to climate change factors while simultaneously regarding advances in analytical methods and clinical diagnostics
(Neil and Berkelman 2008, Ng et al. 2008a). Although Legionella diagnostics have improved and therefore could contribute to increased reporting of the disease (Benin et al. 2002), several studies have shown associations between cases of LD and environmental factors. A case crossover analysis of LD cases in the greater
Philadelphia metropolitan area showed an increase in cases with increased precipitation and humidity (but not temperature), with most cases occurring during the summer months (Fisman 2005). In a similar analysis on national surveillance data from 2003-2007 in the Netherlands, multivariate models showed that mean weekly precipitation intensity, mean weekly temperature, and mean RH accounted for 43.3% of variability in the incidence of LD (Karagiannis et al. 2009). Linear models demonstrated that low sunshine, high cloud cover, mean temperatures close to 17.5°C, high RH, and intense precipitation were associated with higher incidence of LD
(Karagiannis et al. 2009). The hottest days “did not coincide with the highest incidence of LD throughout the year” (Karagiannis et al. 2009).
A study of five U.S. Mid-Atlantic states revealed a 1-cm increase in rainfall was associated with a 2.6% increase in LD incidence; when rainfall increased 5.3 cm from the 1990-2002 summer period to the 2003 summer period, this was associated with a
14.6% increase in LD risk (Hicks et al. 2007). Ricketts et al. (2009) reported an association between RH and LD cases in England and Wales from July to September 23
(2003-2006) but not during winter, with stronger association when maximum temperature was ≥ 20°C. No association was found for wind speed. A summer and autumn seasonal peak of LD is reported in several studies (Fisman 2007, Marston et al. 1994, Ricketts et al. 2009). However, this association may be altered depending on case classification. In a Scotland study, the summer/autumn peak was observed for travel-related cases but an autumn and early winter peak was noted for non-travel cases from 1978 to 1982 and 1983 to 1986, respectively (Bhopal and Fallon 1991).
Nosocomial cases were clustered during the winter months (Bhopal and Fallon 1991).
Decreases in weather events have been offered as an explanation for a decrease in LD in the Netherlands (Euser et al. 2012).
Other variables may also play a role in disease transmission. Several studies have identified thermal inversions as an important factor in the spread of the bacteria over distances as far as 8 km (Addiss et al. 1989, Engelhart et al. 2008, García-Fulgueiras et al. 2003, Pastoris et al. 1997). (Ng et al. 2008b) reported that changes in local hydrology such as low watershed levels (OR=3.6 (95% CI 2.4-5.3) contributed more strongly to LD risk in Toronto, Canada compared to decreased lake temperature (33% increase (95% CI 8-64%) or weather factors like humidity (34% increase, 95% CI 14-
57%) using a case-crossover design.
2.2.1.5. Research gaps
The reviewed literature strongly suggests that the occurrence of Legionella and cases of LD will be impacted by global climate change. Areas for future investigation include:
24
Further characterization of the relationship between LD outbreaks and
meteorological/environmental factors:
o Hydrologic factors (watershed characteristics)
o Extreme weather events (especially precipitation, flooding)
o Changes in natural water environments and implications for entry into
engineered water systems
o Effects of co-contaminants in water/soil
o Atmospheric changes (Temperature inversions, ENSO cycles)
Exploration of development factors that will impact exposure to Legionella
including increased installation and need for maintenance of cooling and
ventilation systems, increased stress on ageing systems in built environments,
use of intermittent and/or vacation housing, and characterization of risks
associated with energy/water efficient devices
Evaluation of Water Safety Plans for mitigating legionellosis risks
Increased surveillance for developing countries and high climate-risk regions
Development of methods for uniform LD surveillance and reporting across
public health systems/regions
25
3. Screening study of Brisbane roof-harvested rainwater tanks for opportunistic pathogens1
3.1. Abstract
A study of six potential opportunistic pathogens (Acanthamoeba spp., Legionella spp.,
Legionella longbeachae, Pseudomonas aeruginosa, Mycobacterium avium and
Mycobacterium intracellulare) and an accidental human pathogen (Legionella pneumophila) in 134 roof-harvested rainwater (RHRW) tanks was conducted using quantitative PCR
(qPCR). All five opportunistic pathogens and accidental pathogen L. pneumophila were detected in rainwater tanks except Legionella longbeachae. Concentrations ranged up to 3.1 x
106 gene copies per L rainwater for Legionella spp., 9.6 x 105 gene copies per L for P. aeruginosa, 6.8 x 105 gene copies per L for M. intracellulare, 6.6 x 105 gene copies per L for
Acanthamoeba spp., 1.1 x 105 gene copies per L for M avium, and 9.8 x 103 gene copies per L for L. pneumophila. Among the organisms tested, Legionella spp. (99% tanks) were the most prevalent followed by M. intracellulare (78%). A survey of tank-owners provided data on rainwater end-uses. Fecal indicator bacteria (FIB) Escherichia coli and Enterococcus spp. were enumerated using culture-based methods, and assessed for correlations with opportunistic pathogens and L. pneumophila tested in this study. Opportunistic pathogens did not correlate well with FIB except E. coli vs. Legionella spp. (tau = 0.151, P = 0.009) and E. coli vs. M. intracellulare (tau = 0.14, P = 0.015). However, M. avium weakly correlated with both L. pneumophila (Kendall’s tau = 0.017, P = 0.006) and M. intracellulare (tau = 0.088, P
= 0.027), and Legionella spp. also weakly correlated with M. intracellulare (tau = 0.128, P =
0.028). The presence of these potential opportunistic pathogens in tank water may present health risks from both the potable and non-potable uses documented from the current survey data.
1 This chapter is published: Hamilton, K., Ahmed, W., Palmer, A., Sidhu, J., Hodgers, L., Toze, S. and Haas, C. (2016) Public health implications of Acanthamoeba and multiple potential opportunistic pathogens in roof-harvested rainwater tanks. Environmental Research 150, 320-327. 26
3.2. Keywords:
Roof-harvested rainwater, opportunistic pathogens, fecal indicator bacteria, quantitative PCR, health risks, rainwater survey
3.3. Introduction
Increasing water scarcity has led to greater reliance on alternative and decentralized potable and non-potable water resources in recent decades (Hanjra et al. 2012). Australia is the driest inhabited continent on Earth and suffered from a severe “millennium” drought from
2001 to 2009 (van Dijk et al. 2013). As a result of water scarcity in this region, the use of roof-harvested rainwater (RHRW) for domestic purposes is a widely accepted practice. This is beneficial for simultaneously conserving water and reducing stormwater runoff.
Pathogens could be introduced to tanks via roof runoff containing fecal matter from birds, insects, bats, possums and reptiles. The microbiological quality of RHRW stored in tanks is generally assessed by monitoring Escherichia coli (E. coli) and Enterococcus spp., which are commonly found in the gut of warm-blooded animals (Albrechtsen 2002, Lee et al. 2010).
The presence of E. coli in tank water generally indicates fecal contamination and the presence of pathogens. Drinking water guidelines have been used to assess the microbial quality of the tank water. For most guidelines, this entails the non-detection of E. coli in 100 mL of water
(NHMRC-NRMMC 2004, WHO 2004). Fecal indicator bacteria should be able to predict human health outcomes. From a public health perspective, the relationship between fecal indicator bacteria and pathogens is critical.
Although case-control studies have established associations between untreated rainwater consumption and gastroenteritis (Brodribb et al. 1995, Merritt et al. 1999), epidemiological studies have not supported a strong linkage (Heyworth et al. 2006, Rodrigo et al. 2011). The presence of multiple non-gastroenteritis-associated microbial pathogens (opportunistic in nature) in rainwater tanks have been reported (Chidamba and Korsten 2015, Dobrowsky et al.
27
2014, Tuffley 1980), supporting the need to assess potential health risks. Only a few studies
(Ahmed et al. 2014a, Kobayashi et al. 2014) have quantified selected numbers of opportunistic pathogens such as Aeromonas hydrophila, Legionella spp., and Staphylococcus aureus by analyzing small numbers of tank water samples.
Opportunistic pathogens infrequently cause illnesses in healthy individuals, and primarily affect those with weakened immune systems, children, and/or the elderly. These pathogens include Legionella spp., Mycobacterium avium complex (MAC), a group of related bacteria that includes both Mycobacterium avium and Mycobacterium intracellulare), Pseudomonas aeruginosa, and Acanthamoeba spp. Acanthamoeba spp. are both pathogens and hosts for other opportunistic bacteria, potentially enhancing their growth and virulence (Falkinham 3rd et al. 2015, Thomas and Ashbolt 2010, Thomas et al. 2010).
Legionella spp. are ubiquitous in water sources and include L. pneumophila, an accidental human pathogen (Mekkour et al. 2013), the most common causative agent of the severe pneumonia-like illness Legionnaires’ Disease, as well as the less severe form, Pontiac fever
(Diederen 2008). Legionellosis is the only disease associated with this group of pathogens that is a nationally notifiable disease in Australia (Australian Government Department of
Health 2016a).The rate of Legionellosis in Australia was 13 people per million people in
2012, which was higher than rates reported for Europe (9.2 people per million), the United
States (10.8 people per million), Canada (4 people per million), Japan (2-7 people per million for 2005 – 2009), and Singapore (6.5 people per million) but lower than New Zealand (14 people per million) (Milton et al. 2012, Phin et al. 2014). However, these estimates are not likely to be directly comparable due to varied approaches for case definition, clinical diagnostics and reporting practices (Phin et al. 2014). Eighty cases of Legionellosis were reported in Queensland, Australia in 2015 (rate of 1.7 per 100,000) (Australian Government
Department of Health 2016b). L. longbeachae has also been isolated from soils and composts linked to outbreaks and is more common in Australia than other countries (Pravinkumar 2010,
Steele 1990). However, the openings in the top of tanks by which water enters are covered with only mosquito mesh, leaving space for dust, soil, fecal matter, on-site compost piles, or
28
other debris to enter the tank. This is hypothesized as a method for potentially introducing L. longbeachae. P. aeruginosa is associated with bacteremia in immunocompromised patients, pneumonia in cystic fibrosis patients, and community-acquired ear, eye, and skin infections, however, systematic information regarding the P. aeruginosa disease burden is not available in Australia (Falkinham 3rd et al. 2015). MAC causes soft tissue infections and cervical lymphadenitis in immune-competent patients and disseminated infections in immune- compromised patients (Falkinham III 1996). Additional known risk factors for MAC are environmental exposures to aerosols containing the bacteria (hot tubs, soil dusts), and host characteristics; in particular, taller, slender post-menopausal women are affected more often by NTM infections than their demographically matched controls (Kartalija et al. 2013). MAC are the most common source of bacterial infection in AIDS patients (Kunimoto et al. 2003,
O'Brien et al. 2000), and most frequently identified isolate in non-tuberculosis mycobacteria
(NTM) cases in the Northern Territory, Australia during 1989 – 1997 (O'Brien et al. 2000).
However, M. intracellulare was identified as the most common pathogen in NTM isolates in
Queensland in 2005 (Thomson et al. 2013b). In the Northern Territory, the yearly incidence of pulmonary MAC disease not associated with human immunodeficiency virus infection was
21 cases per million people in 1997, while in Queensland, the incidence of MAC cases has increased over time from 6.3 per million people in 1985 to 32 per million people in 2005
(O'Brien et al. 2000, Thomson et al. 2013b).
Given the increasing importance of these opportunistic pathogens in municipal drinking water systems, and as a portion of the Australian and global waterborne disease burdens
(Falkinham et al. 2001a, Falkinham et al. 2015, Rusin et al. 1997, Schoen and Ashbolt 2011) it is warranted to further explore their occurrence in RHRW tanks. To the author’s knowledge, only two other studies have measured Mycobacterium spp. (M. intracellulare, M. avium, M. gordonae, and M. terrae-M. triviale-M. nonchromogenicum complex) in RHRW and none to date have quantified these opportunistic pathogens in tank water (Albrechtsen
2002, Tuffley 1980).
29
Culture-based methods have historically been preferred in microbial water quality monitoring efforts. However, their application can be challenging and time-intensive due to the strict growth requirements of certain pathogens, and failure to detect viable but non- culturable (VBNC) pathogens (Bonetta et al. 2010). For example, samples containing
Legionella spp. can take up to 96 h to grow on BCYE agar, and require visual examination of colonies, and further biochemical testing for accurate identification (CDC 2005). In contrast, qPCR is a more rapid and sensitive method for quantifying pathogens that are difficult to grow using culture-based methods. In this study, qPCR methods were chosen for the quantification of seven opportunistic pathogens and culture-based methods were utilized for the enumeration of fecal indicator bacteria (FIB) Escherichia coli and Enterococcus spp.
Although FIB have historically been used as regulatory water monitoring standards for
RHRW, little is known regarding their correlations with opportunistic pathogens in tank water. The aims of this study are therefore to (1) quantify seven potential opportunistic pathogens of public health significance in tank water; (2) assess their correlations with FIB, and (3) highlight the implications of the presence of these opportunistic pathogens in tank water. The quantitative data presented in this study would aid in the quantitative microbial risk assessment (QMRA) of RHRW for various domestic uses.
3.4. Materials and methods
3.4.1. Study areas and survey
Residents were recruited by letter or email (Appendix Figure 9.1) to participate in the study. 134 rainwater tanks were sampled from various areas of Brisbane (n = 84) and the
Currumbin Ecovillage (n = 50), both located in Southeast Queensland (SEQ), Australia during March to September 2015. The Ecovillage is a decentralized residential development that employs a range of strategies to conserve water and energy including a cluster-scale sewage treatment/water reclamation plant, rainwater storage tanks, solar panels, and source-
30
separated urine usage (Hood et al. 2009). Rainwater tank owners were sent an online survey
(prior two days sample collection) regarding end uses (drinking, clothes laundering, car washing, gardening, swimming pool use), and treatment practices (Appendix Figure 9.2). 121
(90%) tank owners provided responses to the survey. On the site, a sanitary inspection was undertaken during sampling to identify factors (the presence of overhanging trees, TV aerials, and wildlife fecal contamination on the roof), and to verify survey results. Results were distributed to residents via email using a standardized template and explanation of all results
(Appendix Figure 9.3- Figure 9.4).
3.4.2. Tank water sampling
The tap/spigot connected directly to the rainwater tank (n = 129) was wiped with 70% ethanol, and the water was run for 30-60 s before filling a 10 L sterile container. In the absence of a tap, sample was collected directly from the opening in the top of the tank (n = 5).
Samples were transported to the laboratory, kept at 4°C, and processed within 6-12 h.
3.4.3. Enumeration of fecal indicator bacteria (FIB)
Colilert® and Enterolert® (IDEXX Laboratories, Westbrook, Maine, USA) Test kits were used to determine the concentrations of FIB (total coliforms, E. coli, and Enterococcus spp. in 100 mL of each tank water sample. Test kits were incubated at 37 ± 0.5°C (for total coliforms and E. coli) and 41.5 ± 0.5°C (for Enterococcus spp.) for 18-24 h as per the manufacturer’s recommendation.
3.4.4. Concentration of rainwater samples
Approximately 10 L water sample from each rainwater tank was concentrated by a hollow-fiber ultrafiltration system (HFUF) using Hemoflow FX 80 dialysis filters (Fresenius
Medical Care, Bad Homberg, Germany) as previously described (Hill et al. 2007, Hill et al.
2005). The sample was concentrated to approximately 300-400 mL and stored at -4°C. A new filter cartridge was used for each sample. The concentrate was filtered through a 0.45 µm
31
cellulose filter paper (Advantec, Tokyo, Japan), and stored at -80°C until DNA extraction. In the case of filter clogging, multiple filter papers were used for each sample. Method blank runs were performed to ensure that the disinfection procedure was effective in preventing carryover contamination between sampling events.
3.4.5. DNA extraction
DNA was extracted using a PowerMax® Soil DNA Kit (Mo Bio Laboratories, Carlsbad,
California, USA) according to the manufacturer’s instructions and stored at -80°C until use.
The kit was modified slightly with 2 mL of DNA eluted buffer C6 instead of 5 mL. DNA concentrations were determined using NanoDrop spectrophotometer (ND-1000, NanoDrop
Technology). Each DNA sample was amplified using a 16S rRNA general bacterial real-time
PCR assay to confirm successful DNA extraction (Boon et al. 2003). All samples tested in this study gave PCR positive and correct amplification (determined by the melt curve analysis) for the 16S rRNA gene of general bacteria.
3.4.6. PCR inhibition
An experiment was conducted to determine the effect of possible PCR inhibitory substances on the quantitative detection of target opportunistic pathogens in tank water DNA samples using a Sketa22 real-time assay (Ahmed et al. 2015). Of the 134 samples, 17% had the sign of PCR inhibition. These inhibited samples were 10-fold serially diluted and further tested with the Sketa22 real-time PCR assay. The results indicated the relief of PCR inhibition at the 10-fold dilution. Based on the results, neat DNA samples (PCR uninhibited samples) and 10-fold diluted (PCR inhibited) samples were tested with qPCR methods.
3.4.7. Preparation of qPCR standards
Standards for each qPCR assay were designed using IDT custom gene synthesis to construct plasmids inserted with a gene fragment (Appendix Table 9.4) from the known sequence (IDTDNA.com). Purified recombinant plasmids containing appropriate base pairs
32
for each qPCR target were produced by Integrated DNA Technologies (pIDTSmart with ampicillin; IDT), and cloned into a vector followed by plasmid extraction (IDTDNA.com;
Coralville, IA). The purified recombinant plasmids were serially diluted to create a standard ranging from 1 × 106 to 1 copies per µL of DNA extract was prepared from the synthesized plasmid DNA. A 3-uL template from each serial dilution was used to prepare a standard curve for each qPCR assay. For each standard, the genomic copies were plotted against the cycle number at which the fluorescence signal increased above the quantification cycle value (Cq value). The amplification efficiency (E) was determined by analysis of the standards and was estimated from the slope of the standard curve as E = 10-1/slope.
3.4.8. qPCR assays
qPCR assays were performed using previously published primers, probes, and optimized reaction mixtures and cycling parameters (Appendix Table 9.5). It is noted that the qPCR assay designed specifically for M. intracellulare developed by Chern et al., 2015 may also quantify M. chimaera; the 16S rRNA sequence for these two microorganisms differs by only one base pair (Chern et al. 2015, Tortoli et al. 2004). All qPCR amplifications were performed in a 20-µL reaction mixture using Sso FastTM Probes Supermix (Bio-Rad
Laboratories, CA). The qPCR mixtures contained 10 µL of Supermixes, optimized concentrations of primers and probe and 3 µL of template DNA. Standards (positive controls) and sterile water (negative controls) were included in each qPCR run. All qPCR reactions were performed in triplicate using a Bio-Rad® CFX96 thermal cycler. qPCR standards were analyzed to determine the amplification efficiencies (E) and the correlation coefficient (r2).
The qPCR lower limit of quantification (LLOQ) was also determined from the Cq values obtained for each standard. The minimum concentration of gene copies from the standard series detected in 100% triplicate samples was considered qPCR LLOQ. Also, to prevent
DNA carryover contamination, reagent blanks were included for each batch of DNA samples.
33
No carryover contamination was observed. To minimize qPCR contamination, DNA extraction and qPCR setup were performed in separate laboratories.
3.4.9. Recovery efficiency
The recovery efficiency of the water sample concentration method was determined using L. pneumophila (ATCC 33152). Purified L. pneumophila were inoculated into BYE broth, and incubated at 37°C overnight. 500 µL of L. pneumophila culture was centrifuged to obtain pellets. The bacterial pellet was washed twice in sterile phosphate buffer saline (PBS) and resuspended in 500 µL PBS. DNA was extracted from the 500 µL of PBS suspensions using a PowerMax® Soil DNA Kit (Mo Bio Laboratories, Carlsbad, California, USA). The concentrations of L. pneumophila in 3 µL DNA samples were determined using qPCR used in this study. 500 µL of bacterial suspension was seeded into 10 L autoclaved rainwater samples
(n = 3). Seeded water samples were concentrated by the HFUF system (the same way the tank water sample was processed). DNA was extracted and using a PowerMax® Soil DNA Kit (Mo
Bio, Carlsbad, California, USA). The concentrations of L. pneumophila in seeded samples were determined using qPCR assay. The recovery efficiency of L. pneumophila was 84% ±
32% (Mean ± SD) in the rainwater samples. Recovery efficiency was calculated as follows:
Recovery efficiency (%) = (concentration recovered/concentration seeded) × 100.
3.4.10. Statistical analysis
All pathogen datasets involved censored data with non-detect or non-quantifiable values [either between the lower limit of detection (LLOD) and lower limit of quantification
(LLOQ) or above the upper limit of quantification (ULOQ),]. In this study, only some culture-based values were > ULOQ. The use of statistical procedures for censored data is necessary to avoid biasing parameter estimates using other non-detect substitution methods
(Helsel, 2011). The Shapiro-Wilk test for normality was used to test the normality of both the raw and log-transformed pathogen data, including non-detect values. The null hypothesis that the data was normally distributed was rejected in both cases (P < 0.05). Therefore,
34
nonparametric Kendall’s Tau correlations were computed to assess relationships between opportunistic pathogens and FIB using the NADA package in the R software environment
(www.r-project.org). Correlations were considered significant for P < 0.05. A Wilcoxon-
Mann-Whitney test for interval-censored data was performed using the interval package (Fay and Shaw 2010) for R to determine the differences between concentrations obtained from
Brisbane area and Currumbin Ecovillage samples. Differences were considered significant for
P < 0.05. Data were plotted using GraphPad Prism version 6.07 (GraphPad Software Inc., La
Jolla, USA).
3.5. Results
3.5.1. Survey data
Ninety-percent of participants responded to the rainwater survey (121 of 134 tanks sampled). 100% of Currumbin respondents (n = 42) used their rainwater for potable use
(Table 3.1). Clothes-washing with rainwater was also common in Currumbin (93%). Non- potable uses were more common for the Brisbane group, with a much lower proportion of respondents from the Brisbane area (n = 75) using their tank water for drinking (20%), cooking (11%), or showering (12%). Gardening was the dominant usage overall (72%), and was more frequently reported in Brisbane (92%) compared to Currumbin (36%). Toilet flushing and pool top-up with rainwater was more commonly reported in Brisbane (33% for each) compared to Currumbin (14% and 2%, respectively). Rainwater was used for car washing comparatively between both Currumbin (41%) and Brisbane (44%). Other less common uses reported included ornamental water features (two outdoor ponds), fish-tank top- up, and pet hygiene.
Currumbin tanks were larger than those from Brisbane (Appendix Table 9.6). Brisbane tanks were most commonly made from polyethylene (68%) and Currumbin tanks were primarily galvanized metal (83%). Roof materials were typically metal or clay/concrete tile.
35
All roofs from Currumbin were metal. Brisbane roofs were more likely to have overhanging trees present (17%) compared to Currumbin (7%). Also, evidence of wildlife droppings on roofs was reported more frequently for Brisbane (63%) than Currumbin (43 %).
36
Table 3.1 Water uses reported by Currumbin (n = 45) and Brisbane (n = 76) rainwater survey participants (one survey entry per tank). Note that multiple uses were reported for each tank and percentages were calculated using the total number of study participants that responded to the question in each location (n = 76 and n = 44, for Currumbin and Brisbane, respectively). The survey was undertaken during March-September 2015.
Location Uses (%)
Drinking Cooking Clothes washing Showering Pool top-up Gardening Car washing Ornamental water Toilet Fish tanks Pet feature flushing filling washing
Currumbin (n = 42 (100) 42 (100) 39 (93) 42 (100) 1 (2.4) 15 (36) 17 (41) 2 (4.8) 6 (14) 0 (0) 0 (0) 42)* Brisbane (n = 75)* 15 (20) 8 (11) 19 (25) 9 (12) 25 (33) 69 (92) 33 (44) 0 (0) 25 (33) 2 (2.7) 2 (2.7)
Total (n = 117)* 57 (49) 50 (43) 58 (50) 51 (44) 26 (22) 84 (72) 50 (43) 2 (1.7) 31 (27) 2 (1.7) 2 (1.7)
*Sample numbers indicate number of participants who responded to the survey in each location and indicated a response to the rainwater usage question 37
3.5.2. Fecal indicator bacteria (FIB)
Of the 84 tank water samples tested from Brisbane, 98%, 68%, and 66% were positive for total coliforms, E. coli, and Enterococcus spp., respectively (Table 3.2). Of the 50 tank water samples tested from Currumbin Ecovillage, 94%, 68%, and 58% were positive for total coliforms, E. coli, and Enterococcus spp., respectively. Among the 134 samples tested, only seven (5%) tanks were free of all three FIB measured. Sixty-six (49%) tank water samples contained both E. coli and Enterococcus spp., and 111 (83%) contained either E. coli or
Enterococcus spp. Concentrations of FIB in positive samples are shown in Figure 3.1.
Concentrations of FIB ranged from 1 to > 2420 MPN per 100 mL for both locations, except
E. coli in Currumbin, which ranged from 1 to 435 MPN per 100 mL. FIB concentrations did not differ significantly for total coliforms, E. coli, or Enterococcus spp. between the Brisbane and Currumbin areas (P = 0.61, P = 0.31, and P = 0.08, respectively). 38
Table 3.2 Occurrence of fecal indicator bacteria (FIB) and potential opportunistic pathogens in Brisbane (n = 84) and Currumbin (n = 50) rainwater samples
Targets No. of positive samples (%)* Brisbane (n = 84) Currumbin (n = 50) Fecal indicators Total coliforms 82 (98) 47 (94) E. coli 57 (68) 34 (68) Enterococcus spp. 55 (66) 29 (58) Potential opportunistic pathogens Acanthamoeba spp. 10 (12) 3 (6) Legionella spp. 81 (96) 50 (100) L. longbeachae 0 (0) 0 (0) L. pneumophila 4 (5) 0 (0) Mycobacterium avium 14 (17) 9 (18) Mycobacterium intracellulare 69 (82) 36 (72) Pseudomonas aeruginosa 21 (25) 17 (34) *Percent positives are calculated as (No. samples positive at location)/(total samples from that location)
Figure 3.1 Box and whisker plots of the log concentrations of fecal indicators in positive samples from 134 rainwater tanks. The inner box lines represent the medians while the outer box lines represent the 25th and 75th data percentiles (Interquartile range, IQR), and the whiskers extend to the range. For each microorganism, dotted lines represent the lower limit of quantification of the respective assay. 39
3.5.3. qPCR standards, lower limit of detection (LLOD) and quantification (LLOQ)
qPCR standards were analyzed to determine the slope, amplification efficiencies, and correlation coefficient values. The standards had a linear range of quantification from 1 × 106 to 1 gene copies per µL of DNA extracts. The slope of the standards ranged from -3.567 to -
3.399. The amplification efficiencies ranged from 90.8% to 95.7%, and the correlation coefficient (r2) ranged from 0.968 to 0.998. qPCR performance characteristics for individual assays are shown in Appendix Table 9.7. The lowest amount of diluted gene copies detected in 100% triplicate samples was considered qPCR LLOQ. LLOQ of the qPCR was determined to be 30 gene copies for Legionella spp., Acanthamoeba spp., and P. aeruginosa assays and 3 gene copies for L. longbeachae, L. pneumophila, M. avium, and M. intracellulare assays.
3.5.4. Concentrations of potential opportunistic pathogens in tank water samples
Table 3.2 shows the percentage occurrence values of six opportunistic pathogens and accidental human pathogen L. pneumophila in tank water samples from Brisbane and
Currumbin. Among the seven organisms, L. longbeachae was not detected in any tank water samples from Brisbane or Currumbin Ecovillage. L. pneumophila (5%) was only detected in
Brisbane samples. Of the 134 water samples, 133 (99%) water samples contained, at least, one opportunistic pathogen, 120 (90%) tanks contained two or more opportunistic pathogens and 52 (39%) tanks contained three or more.
Concentrations of opportunistic pathogens in positive samples are shown in Figure 3.2. The concentrations of Legionella spp. in positive Brisbane samples ranged from 3.0 × 102 to 1.3 ×
105 gene copies per 100 mL in positive samples, followed by P. aeruginosa (3.7 × 102 to 9.6
× 104 gene copies per 100 mL), M. intracellulare (2.3 × 101 to 6.8 × 104 gene copies per 100 mL), Acanthamoeba spp. (2.1 × 102 to 6.6 × 104 gene copies per 100 mL), M. avium (2.2 ×
101 to 1.1 × 104 gene copies per 100 mL), and L. pneumophila (2.4 × 102 to 9.8 × 102 gene 40
copies per 100 mL). The concentrations of Legionella spp. in Currumbin positive samples ranged from 4.0 × 102 to 3.1 × 105 gene copies per 100 mL, followed by P. aeruginosa (3.1 ×
102 to 4.5 × 104 gene copies per 100 mL), Acanthamoeba spp. (2.1 × 102 to 1.8 × 103 gene copies per 100 mL), M. intracellulare (3.3 × 101 to 1.1 × 104 gene copies per 100 mL), M. avium (2.4 × 101 to 4.0 × 102 gene copies per 100 mL). No significant differences in the distribution of potential opportunistic pathogens (all six opportunistic pathogens except L. pneumophila) were observed between Brisbane and Currumbin samples (P > 0.05).
41
Figure 3.2 Box and whisker plots of the log gene copy concentrations of positive samples for each opportunistic pathogen in 134 rainwater tanks. The inner box lines represent the medians, while the outer box lines represent the 25th and 75th data percentiles (Interquartile range, IQR), and the whiskers extend to the range. For each microorganism, dotted lines represent the lower limit of quantification of the respective assay. 42
3.5.6. Correlations among FIB and opportunistic pathogens
Correlations among opportunistic pathogens and between opportunistic pathogens and
FIB are summarized in Table 3.3. All pairs of FIB were significantly correlated. E. coli was significantly associated with Legionella spp. (tau = 0.151, P = 0.009) and M. intracellulare
(tau = 0.14, P = 0.015), while Enterococcus spp. was inversely correlated with M. avium (tau
= -0.092, P = 0.017). Total coliforms were significantly correlated with Legionella spp. (tau =
0.127, P = 0.029) and M. intracellulare (tau = 0.119, P = 0.041). Among the opportunistic pathogens, M. avium was correlated with L. pneumophila (tau = 0.017, P = 0.006) and M. intracellulare (tau = 0.088, P = 0.027) while M. intracellulare was also correlated with
Legionella spp. (tau = 0.128, P = 0.028). Although these correlations were significant, they were weak. 43
Table 3.3 Correlations between opportunistic pathogens and fecal indicators in roof-harvested rainwater. Significant (P < 0.05) values are bold-faced.
Pathogen [Kendall’s tau, (P)]a E. coli Enterococcus spp. Total coliforms Acanthamoeba spp. Legionella spp. L. pneumophila M. avium M. intracellulare Enterococcus spp. 0.252 (<0.001) Total coliforms 0.396 (<0.001) 0.324 (<0.001) Acanthamoeba spp. 0.022 (0.569) 0.059 (0.124) 0.029 (0.454) Legionella spp. 0.151 (0.009) -0.028 (0.628) 0.127 (0.029) -0.007 (0.854) L. pneumophila -0.020 (0.134) 0.005 (0.739) -0.004 (0.760) 0.008 (0.167) -0.026 (0.061) M. avium -0.034 (0.388) -0.092 (0.017) 0.018 (0.655) 0.014 (0.599) 0.048 (0.232) 0.017 (0.006) M. intracellulare 0.140 (0.015) 0.100 (0.080) 0.119 (0.041) 0.072 (0.064) 0.128 (0.028) -0.011 (0.423) 0.088 (0.027) P. aeruginosa -0.009 (0.854) 0.036 (0.430) 0.031 (0.500) 0.009 (0.774) 0.031 (0.506) <0.001 (>0.99) -0.030 (0.343) 0.840 (0.069) a Kendall’s tau is a nonparametric correlation coefficient measuring the monotonic association between y and x. Note that Kendall’s tau is typically approximately 0.15 lower than Spearman’s ρ and Pearson’s r for the same strength of association. 44
3.6. Discussion
Opportunistic pathogens are of increasing public health importance and have been commonly observed in municipal water systems and water storage tanks in association with drinking water biofilms and sediments (Lu et al. 2015, Wang et al. 2014b). These conditions are likely to be comparable to those in rainwater storage tanks, which can receive direct sunlight as well as nutrients from roof-runoff that may be conducive to biofilm growth and, therefore, pathogen proliferation. However, determining factors which may have attributed the opportunistic pathogen proliferation in tank water was not within the scope of this study and warrant further investigation.
Previously, Ahmed et al., (2014) measured four opportunistic pathogens (Aeromonas hydrophila, Staphylococcus aureus, P. aeruginosa and L. pneumophila) in 72 rainwater tank samples within the same geographical location from May to July 2012. Different tanks were sampled during the current study. In addition, the equivalent sample volume used in this study
(10 L) is higher than that reported by previous study where a small volume (1 L) was used for pathogen detection (Ahmed et al. 2014a). Although the current study allowed for increased sensitivity compared to the previous study, the LLOD ranged from 1 to 3 gene copies per µL
DNA which translates to approximately 67- 200 gene copies need to be present in per L for qPCR detection. This value is comparable with the International Organization for
Standardization (ISO) Method ISO method ISO/DIS 11731 detection limit of 50 CFU per L for enumeration of Legionella (ISO 1998, Lehtola et al. 2007). A high recovery efficiency (84
± 32%) was observed in this study, which is comparable with expected recoveries of 70- 93% for the HFUF method reported by Hill et al., (2007).
Ninety-percent of tanks contained two or more opportunistic pathogens. High concentrations (103 - 105 gene copies per L) were detected of all pathogens, except L. longbeachae, which was not detected in any tank water samples tested. Legionalla spp. and
M. intracellulare were the most commonly detected organisms. L. pneumophila was detected at a low rate (3%, up to 9.8 × 103 gene copies per L), which is a lower prevalence but 45
comparable concentrations to those previously reported (up to 1 × 103 gene copies per L)
(Ahmed et al. 2014a, Albrechtsen 2002, Broadhead et al. 1988). L. pneumophila was only detected in four tanks from the Brisbane area and none from Currumbin Ecovillage. Two of the tanks were concrete, and the other two received direct sunlight 75% of the time. Both elevated temperature and surface material roughness are known to affect biofilm growth and therefore the occurrence of opportunistic pathogens; for example, concrete is more prone to biofilm growth than polyvinyl chloride (PVC) water distribution pipes (Hauer 2010, Zhang et al. 2012). Three of four of the L. pneumophila positive tanks had a clay roof and no first-flush device. The tanks had either never been cleaned or cleaned three or more years ago. These factors may have attributed the prevalence of L. pneumophila in Brisbane tank water samples.
L. longbeachae, which is mostly reported in the potting soil mix, has played a role in respiratory disease outbreaks especially in Australia (Kumpers et al. 2008, Lindsay et al.
2012). In Australia and New Zealand, the proportion of Legionnaires’ disease cases associated with L. longbeachae (30%) is greater than it is globally (2-7%) (Victor et al. 2002).
Although there is the potential for soil and other organic materials to enter rainwater tanks through the mosquito mesh or gutter debris accumulation, L. longbeachae was not observed in any samples analyzed during this study. Based on the results of the present study, harvested rainwater is not likely to be a contributor to the L. longbeachae-associated Legionnaires’ disease burden in SEQ but may play a role in L. pneumophila transmission. Still, there are over 50 species of Legionella (Diederen 2008), of which 19 species such as L. bozemanii, L. micdadei, L. feeleii, L. dumoffii, L. wadsworthii, and L. anisa have been documented as human pathogens (Muder and Victor 2002). Future work could elucidate the extent of
Legionella spp. diversity in tanks and determine if any other potential Legionella hazards could be present. Based on a comparison of Legionella spp. and L. pneumophila concentrations in tanks, in tanks where L. pneumophila is present, it represents between approximately 5 and 24% of the total Legionella spp. concentration.
P. aeruginosa and M. avium were also detected frequently in rainwater tanks. P. aeruginosa was previously detected in 13% of rainwater samples (n = 72) with concentrations
46
up to 2 × 103 gene copies per L of water (Ahmed et al. 2014a), which is a lower prevalence and concentration compared to that reported in this study (9.6 × 105 gene copies per L). P. aeruginosa is the most common opportunistic pathogen associated with infections acquired from recreational pools, hot tubs, and whirlpools, particularly due to folliculitis and acute otitis externa (Mena and Gerba 2009, Roser et al. 2014). In the current study, rainwater was commonly used in pools, which could serve as an exposure route for this opportunistic pathogen. However, the use of salt-water pools and/or routine chlorine use was not specified in the survey and could serve to mitigate partially exposure to opportunistic pathogens.
The current findings suggest that rainwater storage tanks are a source of high numbers of
MAC, and M. intracellulare was more commonly present than M. avium. The M. intracellulare measurements here may represent the total concentration of both M. intracellulare and M. chimaera, which differ by one base pair, as mentioned in Chern et al.,
2015. M. avium was observed in 20% (n = 205) of Australian rainwater tank samples
(Tuffley 1980) using culture-based methods, which is comparable with the current findings
(20%). M. avium was also observed in 7% of samples in a study of rainwater in Denmark
(Albrechtsen 2002). However, these studies did not provide quantitative data. MAC has been quantified in potable drinking water distribution systems in Australia at up to a concentration of 1 x 106 gene copies per L (Whiley et al. 2014a), which is an order of magnitude higher than the current findings (up to 1.1 × 105 gene copies per L). Tuffley and Holbeche (1980) also observed other NTM species in rainwater tanks that can cause human disease in humans such as M. gordonae, warranting further exploration of other Mycobacteria spp. in tank water samples that could consider to the NTM disease burden in the future.
The route of exposure to M. avium and M. intracellulare is still not fully understood but is likely to be ingestion or inhalation (Falkinham III 2013a). Not all subspecies of M. avium will infect humans (Hibiya et al. 2011). Therefore, additional work is necessary to identify human- relevant subsets of MAC in rainwater tank samples. While M. avium and M. intracellulare have previously been identified as the major species in drinking water and biofilms
(Falkinham et al. 2001a, Vaerewijck et al. 2005), Wallace et al. recently recently reexamined
47
historical waterborne M. intracellulare isolates and found they were actually M. chimaera by rRNA ITS sequencing (Wallace et al. 2013). Furthermore, M. avium subsp. homissuis and M. chimaera are the primary hazards in municipal drinking water systems while M. intracellulare is more likely to originate from soil or other exposure sources (De Groote et al.
2006, Iakhiaeva et al. 2013). The presence of M. intracellulare in rainwater tank samples determined by a specific qPCR assay may, therefore, reflect the introduction of soil and dust particles into tank water, and present a novel waterborne exposure route for M. intracellulare.
Acanthamoeba spp. can cause eye infections (keratitis) and more rarely granulomatous amoebic encephalitis (GAE) (Nwachuku and Gerba 2004) in contact lens users.
Acanthamoeba play an important role in the drinking water environment by contributing to the growth and modulation of virulence in intracellular bacterial pathogens such as
Legionella, P. aeruginosa, and non-tuberculosis Mycobacteria (NTM, including MAC)
(Delafont et al. 2014, Valster et al. 2011). The presence of Acanthamoeba has been shown to correlate with that of Legionella (Lu et al. 2015), and more recently NTM (Delafont et al.
2014). No correlations of opportunistic pathogens with Acanthamoeba were demonstrated in this study. Previous studies noted correlations between opportunistic pathogens and
Acanthamoeba in sediments, biofilms, or end-points of drinking water distribution systems
(Lu et al. 2015, Wang et al. 2014a). However, rainwater tanks in this study typically have outlets greater than 150-300 mm from the bottom of the tank, so samples may not have gathered water from the region of the tank most prone to biofilm growth. Significant but weak correlations between M. avium and M. intracellulare and Legionella were obtained in this study. Similar findings were found in certain household water heaters supplied by municipal drinking water systems (Wang et al. 2012a), which could be reflective of the high ambient temperatures in Queensland over the study period.
Seventy-percent of tank samples exceeded the Australian Drinking Water Guideline of zero E. coli per 100 mL. Slightly more E. coli was observed in tanks than Enterococcus spp., which is contrary to previous findings (Ahmed et al. 2014a). This may be due to the use of chemical substrate-based methods such as Colilert and Enterolert versus traditional membrane
48
filtration culture-based methodology. In a study of Michigan recreational waters, a comparison of Enterolert and membrane filtration methods for Enterococcus spp. were correlated (R2 = 0.62). However, Enterococcus spp. from Enterolert wells were frequently not verifiable in subsequent microbiological testing, indicating some limitations in comparing these methods (Kinzelman et al. 2003). FIB were significantly correlated with Legionella spp. and MAC, however, this trend was not consistent for all FIB. Therefore, this study is also in accordance with findings that FIB monitoring is not likely to be a reliable surrogate for opportunistic pathogens as previously reported (Ahmed et al. 2014a).
3.7. Conclusions
The survey of Queensland residents indicates a high degree of both potable and non- potable uses of rainwater, and, therefore, the potential for exposure to several potential waterborne opportunistic pathogens of high public health importance. The quantitative data on the concentrations of opportunistic pathogens presented here can be used to undertake a
QMRA to determine the most appropriate usages of rainwater, and if any additional treatment would be prudent. Numerous factors can affect the quality of stored rainwater including roof, piping, and storage system materials, system maintenance, geographic and catchment location, meteorological conditions, atmospheric deposition, the presence of wildlife, tree litter, and tank hydraulics. A longitudinal study is necessary to assess the impact of these factors on pathogen occurrence and variability over time.
49
4. Seasonal assessment of opportunistic pathogens in Brisbane roof-harvested rainwater tanks2
4.1. Abstract
A seasonal study on the occurrence of six opportunistic premise plumbing pathogens
(OPPPs) in 24 roof-harvested rainwater (RHRW) tanks repeatedly sampled over six monthly sampling events (n = 144) from August 2015 to March 2016 was conducted using quantitative qPCR. Fecal indicator bacteria (FIB) Escherichia coli (E. coli) and Enterococcus spp. were enumerated using culture-based methods. All tank water samples over the six events were positive for at least one OPPP (Legionella spp., Legionella pneumophila, Mycobacterium avium, Mycobacterium intracellulare, Pseudmonas aeruginosa, or Acanthamoeba spp.) during the entire course of the study. FIB were positively but weakly correlated with P. aeruginosa (E. coli vs. P. aeruginosa τ = 0.090, p = 0.027; Enterococcus spp. vs. P. aeruginosa τ = 0.126, p = 0.002), but not the other OPPPs. FIBs were more prevalent during the wet season than the dry season, and L. pneumophila was only observed during the wet season. However, concentrations of Legionella spp., M. intracellulare, Acanthamoeba spp., and M. avium peaked during the dry season. Correlations were assessed between FIB and
OPPPs with meteorological variables, and it was determined that P. aeruginosa was the only
OPPP positively associated with an increased antecedent dry period, suggesting stagnation time may play a role for the occurrence of this OPPP in tank water. Infection risks may exceed commonly cited benchmarks for uses reported in the rainwater usage survey such as pool top-up, and warrant further exploration through quantitative microbial risk assessment
(QMRA).
2 This chapter is published: Hamilton, K.A., Ahmed, W., Palmer, A., Smith, K., Toze, S., Haas, C.N. (2017) A seasonal assessment of opportunistic premise plumbing pathogens in roof-harvested rainwater tanks. Environmental Science & Technology, in press. 50
4.2. Keywords: Roof-harvested rainwater; opportunistic pathogens; fecal indicator bacteria; quantitative PCR; public health risks
4.3. Introduction
Roof-harvested rainwater (RHRW) is currently being used globally to supplement potable and non-potable water supplies. As of 2010, 32% of Australian and 36% of Queensland households (n = 1,702,200) had a rainwater tank (ABS 2010). From the year 2007-2010,
Australian urban capital cities experienced the greatest increase in the number of installed
RHRW tanks; Brisbane had the largest increase from 18 to 43% (n = 731,200) (ABS 2010).
RHRW tanks are the main source of drinking water for 13.6% Queensland households (ABS
2010).
RHRW stored in tanks are prone to microbial, chemical, and heavy metal contamination
(Abbasi and Abbasi 2011). Pathogens can enter into the tanks through aerosol deposition, plant litter, and animal fecal matter via roof run-off. In addition, the microbial quality of tank water varies with geographical location, climatic conditions, roof and tank maintenance practices, tank hydraulics, and surrounding environment (Thurman 1995, Vialle et al. 2011).
To monitor the microbial quality of tank water, fecal indicator bacteria (FIB) are most commonly used. The presence of FIB such as Escherichia coli (E. coli) in 100 mL tank water sample indicates fecal contamination and the presence of potential pathogens. Monitoring FIB is cost effective and less technical compared to pathogen monitoring, and is the focus of most rainwater quality guidance documents (NHMRC-NRMMC 2004, WHO 2004). However, evidence continues to support that there are a lack of correlations between FIB and pathogens(Ahmed et al. 2014a, Hamilton et al. 2016, Harwood et al. 2005). Therefore, FIB monitoring may not be sufficient to provide information on health risks.
Although case control studies have indicated some associations between untreated rainwater consumption and gastroenteritis (Brodribb et al. 1995, Merritt et al. 1999), epidemiological studies have not supported a strong linkage (Heyworth et al. 2006, Rodrigo et
51
al. 2011). However, these studies mainly focused on enteric pathogens. In recent years, improvements in disinfection practices have reduced the health burden of gastroenteritis- causing pathogens in many centralized drinking water systems. As a result, the focus for waterborne disease burden mitigation has shifted to opportunistic premise plumbing pathogens (OPPPs) that live in the biofilms growing on the inner surfaces of the distribution system, premise plumbing pipes, and bulk water within a household water system (Pruden et al. 2013). OPPPs are frequently measured in drinking water systems, however, rainwater tanks mimic the conditions of a high OPPP-risk system due to their high residence times, potentially high nutrient content, and suitable temperatures, especially in sub-tropical regions.
OPPPs infrequently cause illnesses in healthy individuals, and primarily affect those with weakened immune systems, children, and/or the elderly (Falkinham 3rd et al. 2015). OPPPs include Acanthamoeba spp., Legionella pneumophila, Mycobacterium avium complex (MAC, a group of related bacteria that includes both Mycobacterium avium and Mycobacterium intracellulare), and Pseudomonas aeruginosa, among others (Falkinham 3rd et al. 2015,
Pruden et al. 2013). Acanthamoeba spp. and other protozoans can form relationships with bacteria and contribute to their proliferation and enhanced virulence factors (Thomas et al.
2010). L. pneumophila causes Legionellosis (Legionnaires’ Disease) which is the only reportable OPPP-associated illness in Australia (Australian Government Department of
Health 2016a)and Pontiac Fever. Legionellosis had an estimated incidence of 13 people per million in Australia in 2012 (Phin et al. 2014). Sporadic outbreaks have been linked specifically to L. pneumophila and MAC in people who were exposed to RHRW (Lumb et al.
2004, Schlech III et al. 1985, Simmons et al. 2008).
The presence of multiple OPPPs in tank water samples has been reported by several studies (Ahmed et al. 2014a, Ahmed et al. 2008, Albrechtsen 2002, Chidamba and Korsten
2015, Dobrowsky et al. 2014, Hamilton et al. 2016), supporting the need to assess potential health risks. However, studies that investigated the occurrence of OPPPs relied upon testing a sample at a single time-point from a tank. Therefore, there is limited or no data available on the temporal variations in OPPPs occurrence and concentrations. Information regarding
52
occurrence and temporal variations of OPPPs is particularly important because water quality management options for individual rainwater tank owners are limited. In addition, there are barriers associated with the level of technical knowledge necessary, maintenance, and costs.
In our previous study(Hamilton et al. 2016), we screened 134 roof-harvested rainwater tank samples for OPPPs. Of the 134 tanks, 24 were further selected for this follow-up study.
The aims of this study were therefore to: (i) provide seasonal data on the concentrations of six potential OPPPs (Acanthamoeba spp., Legionella spp., L. pneumophila, M. avium, M. intracellulare, and P. aeruginosa) of public health significance in tank water over time (ii) assess the overall correlations among OPPPs and FIB; and (iii) compare the presence of
OPPPs with maintenance and rainwater system characteristics. qPCR methods were chosen for the quantification of six OPPPs and culture-based methods were used for the enumeration of two FIB. The quantitative data presented in this study would improve the accuracy of quantitative microbial risk assessment (QMRA) of RHRW for various domestic uses, and provide information for rainwater users regarding potential seasonality of risks.
4.4. Materials and Methods
4.4.1. Tank water sampling.
Water samples were collected in two phases. In phase one, a total of 134 water samples were collected from 134 tanks located in various areas of Brisbane and the Currumbin
Ecovillage in Southeast Queensland, Australia between March and July 2015 as part of a previous study (Hamilton et al. 2016) (phase one). In phase 1, the concentrations of FIB (E. coli and Enterococcus spp.) were determined using culture-based methods, and the gene copies for six OPPPs (Acanthamoeba spp., Legionella spp., L. pneumophila, M. avium, M. intracellulare, and P. aeruginosa) were quantified using qPCR assays. Based on the high concentrations of FIB and OPPPs among the 134 tank water samples (Appendix Table 9.8), a subset of tanks with high concentrations of OPPPs and FIB (n = 24) was selected for the current seasonal study (phase two).
53
In phase two, for the seasonal study, monthly water samples were collected from these 24 selected tanks on six separate events from August 2015 to March 2016, giving a total number of 144 tank water samples. The tap/spigot connected directly to the rainwater tank was wiped with 70% ethanol, and the stored water was run for 15 s prior to filling a 10 L sterile container. Samples were immediately transported to the laboratory, kept at 4°C, and processed within 6-12 h.
4.4.2. Tank survey
Tank owners were sent an online survey regarding end uses (drinking, clothes laundering, car washing, gardening, swimming pool use), and treatment practices. On site, a visual sanitary inspection was undertaken to identify factors that may have been associated with sources (the presence of overhanging trees, TV aerials, and wildlife fecal contamination on the roof), and to verify the online survey results provided by the residents. Any additional factors that might contribute to the presence of fecal contamination in the tank water were noted for each property.
4.4.3. Enumeration of FIB
Colilert® and Enterolert® (IDEXX Laboratories, Westbrook, Maine, USA) Test kits were used to determine the concentrations of FIB (total coliforms, E. coli, and Enterococcus spp. in 100 mL of each tank water sample. Test kits were incubated at 37 ± 0.5°C (E. coli) and 41.5 ± 0.5°C (for Enterococcus spp.) for 24 h as per the manufacturer’s recommendation.
4.4.4. Concentration of rainwater samples and DNA extraction
Approximately 10 L water sample from each rainwater tank was concentrated by a hollow-fiber ultrafiltration system (HFUF) using Hemoflow FX 80 dialysis filters (Fresenius
Medical Care, Bad Homburg, Germany) as previously described (Hamilton et al. 2016, Hill et al. 2005). The concentrated sample was filtered through a 0.45 µm cellulose filter paper
(Advantec, Tokyo, Japan), and stored at -80°C until DNA extraction. In case of filter
54
clogging, multiple filter papers were used for each sample. DNA was extracted using a
PowerSoil® Max DNA Kit (Mo Bio, Carlsbad, California, USA) according to the manufacturer’s instructions. The kit was modified slightly with 2 mL of DNA eluted buffer
C6 instead of 5 mL (Hamilton et al. 2016). DNA concentrations were determined using a
NanoDrop spectrophotometer (ND-1000, NanoDrop Technology) and stored at -80°C until use.
4.4.5. PCR inhibition
An experiment was conducted to determine the effect of potential PCR inhibitory substances on the quantitative detection of OPPPs in tank water DNA samples using a
Sketa22 real-time assay (Hamilton et al. 2016, Haugland et al. 2005). Of the 144 samples, 13
(9%) had signs of PCR inhibition. These inhibited samples were 10-fold serially diluted, and further tested with the Sketa22 real-time PCR assay. The results indicated the relief of PCR inhibition at the 10-fold dilution. Based on the results, neat DNA samples (PCR uninhibited samples) and 10-fold diluted (PCR inhibited) samples were tested with qPCR methods.
Samples with a 2 quantification cycle (Cq) delay were considered as having PCR inhibitors.
4.4.6. qPCR standards
Standard curves for all qPCR assays were constructed using synthesized plasmid
DNA (pIDTSMART with ampicillin resistance; Integrated DNA Technologies, Coralville,
IA, USA). The purified plasmid DNA was serially diluted to create a standard ranging from 1
× 106 to 1 gene copies per µL of DNA. A 3-uL template from each serial dilution was used to prepare a standard curve for each qPCR assay. For each standard, the genomic copies were plotted against the cycle number at which the fluorescence signal increased above the quantification cycle value (Cq value). The amplification efficiency (E) was determined by analysis of the standards and was estimated from the slope of the standard curve as E = 10-
1/slope.
55
4.4.7. qPCR assays and performance characteristics
qPCR assays were performed using previously published primers, probes, and optimized reaction mixtures and cycling parameters (Appendix Table 9.9). All qPCR amplifications were performed in a 20-µL reaction mixture using Sso FastTM Probes Supermix
(Bio-Rad Laboratories, CA). The qPCR mixtures contained 10-µL of Supermixes, optimized concentrations of primers and probe and 3-µL of template DNA. Standards (positive controls) and sterile water (negative controls) were included in each qPCR run. All qPCR reactions were performed in triplicate using a Bio-Rad® CFX96 thermal cycler. qPCR standards were analysed in order to determine the amplification efficiencies (E) and the correlation coefficient (r2). The qPCR lower limit of quantification (LLOQ) was also determined from the Cq values obtained for each standard. The minimum concentration of copies from the standard series detected in 3/3 qPCR reactions was considered qPCR LLOQ. The dilution below this series was considered the limit of detection (LOD).
The slope of the standards ranged from -3.567 to -3.399. The amplification efficiencies ranged from 90.8 to 95.7%, and the correlation coefficient (r2) ranged from 0.968 to 0.998. qPCR performance characteristics for individual assays are shown in Appendix
Table 9.10. The lowest amount of diluted gene copies detected in 3/3 qPCR reactions was considered the qPCR LLOQ. The LLOQ of the qPCR was determined to be 30 gene copies for Legionella spp., Acanthamoeba spp., and P. aeruginosa assays and 3 gene copies for L. pneumophila, M. avium, and M. intracellulare assays.
4.4.8. Quality control
Method blank runs were performed to ensure that the disinfection procedure was effective in detecting carryover contamination between sampling events. In addition, to detect
DNA carryover contamination, reagent blanks were included for each batch of DNA samples.
No carryover contamination was observed. To minimize qPCR contamination, DNA extraction and qPCR setup were performed in separate laboratories.
56
4.4.9. Meteorological data
Data for daily and monthly rainfall, ambient temperature, and relative humidity (RH) at 9am and 3pm on the sampling day were obtained from the Australian Bureau of Meteorology
(BOM) (Australian Government Bureau of Meteorology 2016). The closest BOM gauges for daily rainfall data and complete monthly meteorological data were determined by mapping gauges and rainwater tanks and determining the gauges, which were present at the shortest geodesic distance from each tank (Supplemental Table 9.11). Gauges A-L are closest to the
24 sites, and were used for rainfall analysis. Gauges 1-5 were the closest sites that provided full meteorological datasets for analysis of temperature and RH.
4.4.10. Statistical analysis
Statistical analyses were carried out to answer the questions (1) Does FIB/OPPP occurrence differ over the six sampling events? and (2) Is each FIB/OPPP correlated with another FIB/OPPP or other meteorological factors? Differences in OPPP occurrence over six sampling events from August 2015 to March 2016 were assessed using binary presence/absence data as well as their concentrations using the SPSS software (IBM, version
24) and R (www.rproject.org). All samples above the LOD and LOQ were considered positive.
4.4.11. Differences in binary (presence/absence) occurrence across six sampling events
Differences in the number of positive samples for each target over time were determined using a Cochran Q test (Cochran 1954). Cochran’s Q test is a method of testing for differences in frequencies or proportions between three or more repeated measures groups.
Results were considered significant for an alpha level of 0.05. Post-hoc McNemar’s tests were performed where significant differences were indicated, and compared to an adjusted alpha value using the Bonferroni correction (significance level alpha/number of pairwise comparisons), alphaadjusted = 0.05/15 = 0.003.
57
4.4.12. Differences in concentrations of FIB / OPPPs (continuous) occurrence across six sampling events
A Shapiro-Wilk test was used to test the normality of the log-transformed OPPP data for each sampling event. Non-detect observations were substituted with half of the detection limit value for this purpose only. The null hypothesis that the data were normally distributed was rejected in all cases (p < 0.05) with the exception of Legionella spp. in February 2016 (p
= 0.145). Quantile-quantile and residual plots were also examined for normality and agreed with these findings. As a result, the nonparametric Friedman test for repeated measures data was used to determine differences between concentrations of microorganisms obtained during the different sampling events. Results were considered significant for an alpha level of 0.05.
Post-hoc Wilcoxon signed-rank tests for differences in median rank scores were performed where significant differences were indicated and compared to an adjusted alpha value of
0.003 using the Bonferroni correction.
4.4.13. Correlations among FIB/OPPPs and between FIB/OPPPs and meteorological factors
Correlations were assessed among FIB and OPPPs, and between FIB/OPPPs and meteorological parameters using binary presence/absence data as well as FIB and OPPP concentrations. For binary FIB / OPPP data, Odds ratio (OR) estimates and 95% confidence intervals (CI) for the estimates were calculated between FIB and OPPPs for the pooled dataset using a chi-square test. For binary pathogen data vs. meteorological data, binary logistic regression was used to calculate ORs. Fisher's exact test was used to assess the significance of the ORs, where an odds ratio greater than one (with a 95% CI that does not overlap 1) indicates a positive association between two factors while an odds ratio less than one (with a
95% CI that does not overlap 1) indicates a negative association.
To assess correlations among continuous FIB/OPPP concentrations and meteorological information for all sampling events, nonparametric Kendall’s Tau correlations were computed
58
using the NADA package in the R software environment (www.r-project.org). Although a
Spearman Rho is typically used for this purpose, pathogen datasets involving censored data with non-detect (below the method detection limit) and non-quantifiable (below the limit of quantification) values require the use of Kendall’s Tau as the Spearman rho does not accommodate multiple reporting limits for censored datasets (Helsel 2011). Correlations were considered significant at an alpha level of 0.05.
We performed multiple comparisons for both the OR and Kendall’s Tau analysis and used the false discovery rate (FDR) approach to correct our decisions on significance based on the number of comparisons and the p-values of the tests that we performed (Benjamini and
Hochberg 1995). For tests between the 8 microorganisms (6 OPPP + 2 FIB), 28 comparisons were performed and the modified p-value for significance (assuming a 10% FDR and noting that FDR is not equivalent to α(McDonald 2009)) was 0.017 for odds ratios and 0.031 for
Kendall’s Tau analysis. For tests between the 8 microorganisms and 17 meteorological variables, 136 comparisons were performed and the modified p-value for significance (10%
FDR) was 0.016 for odds ratios and 0.021 for Kendall’s Tau analysis.
4.5. Results
4.5.1. Survey data
Ninety-six percent (23 of 24 tanks) of tank owners responded to the survey. The sizes of the tanks ranged from 3,000 to >40,000 L, and the tank ages ranged from 2 to 20 years
(Table 4.1). Most of the tanks were made of galvanized steel (50%) or polyethylene (37.5%), whereas, only three tanks (12.5%) were made with concrete. Most of the houses had metal roof (87.5%) and only 3 houses (12.5%) had tile roofs. Of the 24 tanks surveyed, 21% had trees overhanging the roof, 46% had TV aerials, and 33% had visible signs of debris on the roof. Fifty-percent of the tanks were never cleaned and desludged in their lifetime. Of the 24 tanks, 46% had a first-flush diverter installed and only 38% treated the water (all used
59
filtration) before drinking. 54% tanks were used for both potable and non-potable purposes and the remaining tanks were used only for non-potable purposes.
60
Table 4.1 Characteristics and end uses of rainwater systems tested in this study
Tanks Location Size Age Material Roof material Overhanging TV aerial Wildlife Gutter cleaning Time since last First End uses(s) Treatment (liters) (yr) trees on the on the droppings frequency tank cleaning/de- flush roof roof on roof sludging diverter T1 Periurban >15,000- 5 Galvanized steel Metal N N Y Never Never N Potable, N 20,000 nonpotable T2 Urban 3,000- 8 Polyethylene Metal N Y N Never >3 yr N Nonpotable N 5,000 T3 Urban - 7-8 Polyethylene Metal N N N Never >3 yr N Nonpotable N T4 Periurban >10,000- 7 Polyethylene Metal N N N Yearly >3 yr - Potable, N 15,000 nonpotable T5 Urban >5,000- 6 Polyethylene Metal N Y Y Yearly Never N Nonpotable N 10,000 T6 Urban - 5 Galvanized steel Metal Y Y - Gutter guards Never N Nonpotable N T7 Periurban >5,000- 4 Polyethylene Metal N Y N Never, has gutter >3 yr Y Nonpotable N 10,000 guards T8 Urban 3,000- 5.5 Concrete Clay/ N Y N Yearly Never N Nonpotable N 5,000 concrete tile T9 Urban >5,000- 9 Polyethylene Metal N N N Yearly Never N Nonpotable N 10,000 T10 Urban 3,000- 5 Polyethylene Metal Y N N Yearly >3 yr Y Nonpotable N 5,000 T11 Urban >10,000- 10 Polyethylene Metal N N Y Never >3 yr Y Nonpotable N 15,000 T12 Rural >40,000 2 Galvanized steel Metal N Y N Never Never Y Potable, Y nonpotable T13 Rural - 5 Galvanized steel Metal N N N - >2-3 yr N Potable, Y nonpotable T14 Periurban >40,000 3 Galvanized steel Metal N N N Never Never Y Potable, Y nonpotable T15 Rural >20,000- 7 Concrete Metal Y N N Yearly >3 yr Y Potable, Y 40,000 nonpotable T16 Rural >15,000- 4 Galvanized steel Metal N N Y Never Never N Potable, Y 20,000 nonpotable T17 Periurban >40,000 4 Galvanized steel Metal N Y N Gutter guards Never Y Potable, N nonpotable T18 Periurban >40,000 3 Colourbond Metal N Y Y Every 1.5 yr Never Y Potable, Y Steel nonpotable T19 Periurban >40,000 7 Galvanized steel Metal N N Y Yearly Never Y Potable, Y nonpotable T20 Periurban >40,000 8 Galvanized steel Metal N N N Never >3 yr Y Potable, Y nonpotable T21 Periurban - - Galvanized steel Metal Y - Y - - - Potable, Y nonpotable T22 Rural >15,000- 20 Galvanized steel Metal N Y Y Yearly >3 yr N Potable, N 20,000 nonpotable T23 Urban 3,000- 4 Polyethylene Clay/ N Y N Never Never N Nonpotable N 5,000 concrete tile T24 Periurban >15,000- 8 Concrete Clay/ Y Y N Every 6 months >3 yr Y Nonpotable N 20,000 concrete tile 61
4.5.2. Meteorological data
Queensland is located in the sub-tropical climate zone of Australia, and accordingly exhibits wet (November to March) and dry (April to October) seasons (Wells 2013).
Consistent with this pattern, total monthly rainfall generally increased over the sampling period from August to March, with the exception of February which was unusually dry for the region (Appendix Figure 9.5). Rainfall was generally higher for coastal areas (Gauge L, Gold
Coast) compared to inland samples (Gauge G, urban Brisbane) and was highly localized
(Appendix Figure 9.6). Most sampling events were taken within 48 h of a rainfall event and all were taken within 7 days of a rain event with the exception of nine samples in February.
The February sampling time point was maintained in order to maintain the seasonal sampling scheme every 3-4 weeks. Maximum daily rainfall in Brisbane (Gauge G) ranged from 7.8 mm
(February) to 26.4 mm (March) during the sampling period. Maximum daily rainfall on the
Gold Coast (Gauge L) ranged from 9.2 mm (February) to 98.2 mm (November).
The Queensland wet season is typically accompanied by higher temperatures and humidity. Temperatures increased steadily over the sampling period (Appendix Figure 9.7) with the greatest variations in daily temperature occurring at Gauge 2. Large daily variations in RH over the sampling period are shown in Appendix Figure 9.8. For 100/144 sampling times, RH increased between 9am and 3pm. Average daily RH change calculated as (|RH9am –
RH3pm|/ (No. samples in group) was 9.34% increase (maximum increase 55%) for increasing samples, and 8.02% decrease (max decrease 16%) for decreasing samples.
4.5.3. FIB and OPPPs in tank water samples collected in phase 1
In all, 134 tank water samples were screened for the FIB and OPPPs in phase 1.
Twenty-four tanks were further selected for longitudinal study based on the high occurrence of FIB and OPPPs. The data obtained for only the selected 24 tank water samples is presented. The 24 samples tested were positive for E. coli (63%), Enterococcus spp. (50%),
Acanthamoeba spp. (17%), Legionella spp. (100%), L. pneumophila (17%), M. avium (88%), 62
M. intracellulare (92%), and P. aeruginosa (21%). The concentrations of FIB and OPPPs in tank water samples are shown in Appendix Table 9.8.
4.5.4. FIB and OPPPs in tank water samples collected in phase 2
Among the 144 tank water samples tested in phase 2 (24 tanks × six events), 44, 42, 40,
97, 5, 57, 60, and 31% were positive for E. coli, Enterococcus spp., Acanthamoeba spp.,
Legionella spp., L. pneumophila, M. avium, M. intracellulare, and P. aeruginosa, respectively. Notably, all tank water samples (denoted T1 through T24 for tank 1 through tank 24) over six events were positive for at least one OPPP during the entire course of the study (Table 4.2). All except three tanks for E. coli (T6, T9, T15) and Enterococcus spp. (T5,
T6, T23) were positive at least once during the course of the study. Among the OPPPs tested,
Legionella spp. were detected most frequently in all tank water samples during the course of the study. Acanthamoeba spp. were present at least once in all tanks over the course of the study, with the exception of T14. M. avium (T2, T5, T9, T10, T16, T21) and M. intracellulare
(T2, T3, T5, T9, T12, T13, T16, T19, T20) were frequently detected in several tanks during the course of the study. P. aeruginosa also intermittently detected in small numbers of tank water samples collected in August, September, October and November and then all tank water samples collected in February were PCR positive. In March, P. aeruginosa was detected in over half of the tank water samples. L. pneumophila was only found in three tanks (T8, T22 and T24) and only during the second half of the study (November 2015 through March 2016).
Concentrations of FIB in positive samples are shown in Figure 4.1. Concentrations of E. coli and Enterococcus spp. in positive samples ranged from 1 to 687 and 1 to > 2419 MPN per 100 mL of tank water, respectively. The concentrations of FIB in water samples collected in August, September, October and November were lower than February and March.
Concentrations of OPPPs in positive samples are shown in Figure 4.2. In order of highest to lowest maximum concentration, the concentrations of P. aeruginosa in positive samples were
3.6 × 102 to 4.7 × 108 gene copies per 100 mL, followed by Legionella spp. (3.2 × 102 to 2.3 ×
106 gene copies per 100 mL), Acanthamoeba spp. (2.2 × 102 to 9.8 × 105 gene copies per 100
63
mL), M. intracellulare (2.2 × 101 to 6.8 × 105 gene copies per 100 mL), M. avium (2.4 × 101 to 3.6 × 105 gene copies per 100 mL), and L. pneumophila (2.3 × 101 to 1.5 × 102 gene copies per 100 mL).
64
Table 4.2 Occurrence of fecal indicator bacteria and opportunistic pathogens (OPPPs) in tank water samples (n = 24) over six events
Tanks E. coli Enterococcus spp. Acanthamoeba spp. Legionella spp. L. pneumophila M. avium M. intracellulare P. aeruginosa
A S O N F M A S O N F M A S O N F M A S O N F M A S O N F M A S O N F M A S O N F M A S O N F M T1
T2
T3
T4
T5
T6
T7
T8
T9
T10
T11
T12
T13
T14
T15
T16
T17
T18
T19
T20
T21
T22
T23
T24
TOTAL
7 7 9 5 4 9 9 9 3 6 0 0 0 2 2 3 9 3 2 3 1
13 17 10 14 15 13 22 23 23 23 24 23 23 11 12 24 11 15 16 17 13 17 10 13 24 11 POS 11 A: August; S: September; O: October; N: November; F: February: M: March; ■ represents samples that were above quantification limit; ■ represents samples that are positive but not quantifiable; ■ represents samples that were below detection limit 65
Figure 4.1 Concentrations of fecal indicator bacteria (FIB) in roof-harvested rainwater (RHRW) from Queensland, Australia over six monthly sampling events. Red dotted line denotes the limit of quantification; horizontal hash marks denote the median concentration. Where no horizontal hash marks are present, the median was below the detection limit.
66
Figure 4.2 Concentrations of opportunistic premise plumbing pathogens (OPPPs) in roof-harvested rainwater (RHRW) from Queensland, Australia over six monthly sampling events. Red dotted line denotes the limit of quantification; Green lines denote the limit of detection; horizontal hash marks denote the median concentration. Where no horizontal hash marks are present, the median was below the detection limit. 67
4.5.5. Differences in FIB and OPPPs occurrence across six sampling events
Regarding variation in the proportion of samples positive for FIB and OPPPs over the six sampling events, a Cochran’s Q test determined that there was a significant difference for
E. coli (Q = 14.1, p < 0.05), Enterococcus spp. (Q = 21.7, p < 0.05), Acanthamoeba spp. (Q =
37.7, p < 0.001), M. avium (Q = 34.0, p < 0.001), and P. aeruginosa (Q = 72.7, p < 0.001) over time. However, there was no significant difference for Legionella spp. (Q = 1.09, p >
0.05), M. intracellulare (Q = 11.1, p = 0.05), or L. pneumophila (Q = 10.6, p > 0.05).
Significant post-hoc differences are summarized in Appendix Table 9.12. Notably,
Acanthamoeba spp. were detected significantly more frequently in September than all other months (p <0.001- 0.001), and P. aeruginosa was detected significantly more frequently in
February than in all other months (p < 0.001).
A Friedman test determined that there was a significant difference between the mean ranks over sampling events for E. coli (χ2 = 22.3, p < 0.001), Enterococcus spp. (χ2 = 18.8, p
< 0.05), Legionella spp. (χ2 = 65.3, p < 0.001), Acanthamoeba spp. (χ2 = 39.3, p < 0.001), M. avium (χ2 = 52.9, p < 0.001), M. intracellulare (χ2 = 48.4, p < 0.001), and P. aeruginosa (χ2 =
87.4, p < 0.001). Significant post-hoc differences are summarized in Appendix Table 9.13.
Notably, Acanthamoeba spp. concentrations were significantly higher in September than
February (p <0.001) or March (p <0.001), Legionella spp. concentrations were significantly lower in February and March than all other months (p < 0.001 - 0.001), M. intracellulare concentrations were significantly higher in September than all proceeding months (p < 0.001
– 0.002), and P. aeruginosa concentrations were significantly higher in February than all other months (p <0.001).
4.5.6. Correlations among FIB and OPPPs
Odds ratios (OR) (with 95% confidence intervals) were calculated among FIB and OPPPs
(Figure 4.3). OR estimates among FIB and OPPPs were applied to all datasets collected during the course of the study (August 2015 to March 2016). E. coli and Enterococcus spp. 68
were significantly correlated (OR = 3.2). Enterococcus spp. also had higher odds of occurring when P. aeruginosa (OR = 3.08) was present relative to when it was absent. M. intracellulare had higher odds of occurring when Acanthamoeba spp. (OR = 5.15) or Legionella spp. (OR =
2.62) were present. All these associations were statistically significant and had a p-value <
0.017.
Correlations among FIB and pathogen concentrations are shown in Table 4.3. Note that although this appears to be a weak correlation, Kendall’s tau is typically approximately 0.15 lower than Spearman’s ρ and Pearson’s r for the same strength of association (Helsel and
Hirsch 2002). FIB were weakly but significantly correlated only with P. aeruginosa (E. coli τ
= 0.090, Enterococcus spp. τ = 0.126). E. coli and Enterococcus spp. were correlated (τ =
0.208). Among OPPPs, Acanthamoeba spp. was significantly correlated with M. intracellulare (τ = 0.145) and Legionella spp. (τ = 0.126). Legionella spp. was significantly correlated with M. avium (τ = 0.293), M. intracellulare (τ = 0.260), and inversely correlated with P. aeruginosa (τ = -0.209). M. intracellulare (τ = -0.132) were also inversely correlated with P. aeruginosa.
69
Table 4.3 Correlations between fecal indicator bacteria (FIB) and opportunistic premise plumbing pathogens (OPPPs) in roof-harvested rainwater (RHRW) stored in tanks. Significant values (p < 0.031) are bold-faced; see methods for discussion of correcting for multiple comparisons using a FDR approach.
Pathogen (Kendall’s E. coli Enterococcus Acanthamoeba Legionella spp. L. pneumophila M. avium M. intracellulare tau, p)a spp. spp. Enterococcus spp. 0.208 (<0.001) Acanthamoeba spp. -0.038 (0.399) -0.511 (0.244) Legionella spp. -0.087 (0.087) -0.085 (0.090) 0.126 (0.011) L. pneumophila 0.010 (0.557) 0.028 (0.076) -0.013 (0.407) 0.0003 (0.992) M. avium -0.060 (0.222) -0.077 (0.110) 0.045 (0.341) 0.293 (<0.001) -0.017 (0.335) M. intracellulare -0.011 (0.837) -0.038 (0.434) 0.145 (0.002) 0.260 (<0.001) -0.002 (0.896) 0.046 (0.372) P. aeruginosa 0.090 (0.027) 0.126 (0.002) -0.069 (0.084) -0.209 (<0.001) 0.012 (0.393) -0.093 (0.031) -0.132 (0.003) aKendall’s tau is a nonparametric correlation coefficient measuring the monotonic association between y and x. Note that Kendall’s tau is typically approximately 0.15 lower than Spearman’s ρ and Pearson’s r for the same strength of association (Helsel and Hirsch 2002).
70
Figure 4.3 Odds ratio and 95% confidence intervals for the ability of individual FIB and opportunistic premise plumbing pathogens (OPPPs) to predict the presence of other fecal indicator bacteria (FIB) and OPPPs in pooled tank water samples (n = 144). * indicates significance level (p < 0.017); see methods for discussion of correcting for multiple comparisons using a FDR approach. EC = E. coli, ENT = Enterococcus spp., ACAN = Acanthamoeba spp., LEG = Legionella spp., LP = L. pneumophila, MA = M. avium, MI = M. intracellulare, PA = P. aeruginosa. 71
4.5.7. Correlations between FIB/OPPPs and meteorological factors
Binary logistic regression results between meteorological parameters and the presence or absence FIB and OPPPs are shown in Appendix Table 9.14. The temperature on the sampling day (taken either as maximum or minimum for each day) was significantly positively correlated with the presence of E. coli, Enterococcus spp., L. pneumophila, and P. aeruginosa but negatively correlated with Acanthamoeba spp. (Appendix Table 9.14).
Average maximum monthly temperature was positively and significantly correlated with E. coli (OR = 1.162, Χ2= 6.75, Nagelkerke R2 = 0.06), Enterococcus spp. (OR = 1.245, Χ2= 13.5,
Nagelkerke R2 = 0.12), and P. aeruginosa (OR = 1.692, Χ2= 46.0, Nagelkerke R2 = 0.389), but negatively correlated with Acanthamoeba spp. (OR = 0.828, Χ2= 10.2, Nagelkerke R2 =
0.09). RH at 9 am or 3 pm on the sampling day was negatively correlated with
Acanthamoeba spp.. Total monthly rainfall did not correlate with any microorganism, while rainfall on one or more of the seven days antecedent to the sampling day was positively correlated with the number of samples positive for E. coli, Acanthamoeba spp., M. intracellulare, and negatively correlated with P. aeruginosa. Notably, the time to the last rain event (positive, OR = 1.137, Χ2= 6.6, Nagelkerke R2 = 0.06) and last rainfall event size
(negative, OR= 0.813, Χ2= 15.9, Nagelkerke R2 = 0.15) were correlated only with P. aeruginosa.
Correlations between continuous concentrations of FIB/OPPP and meteorological factors are shown in Appendix Table 9.15. Rainfall on the sampling day or 7 prior days was positively and significantly (but weakly) correlated with E. coli and L. pneumophila concentrations, but negatively (and weakly) correlated with Legionella spp., L. pneumophila,
M. avium, and P. aeruginosa concentrations. Acanthamoeba spp. did not display a consistent trend with rainfall. Average maximum monthly and daily min or max temperatures on the sampling day were positively correlated with FIB and P. aeruginosa, but negatively correlated with OPPPs Acanthamoeba spp., Legionella spp., M. avium, and M. intracellulare.
RH on the sampling day (at 9am and/or 3pm) was positively correlated with Enterococcus 72
spp. and L. pneumophila, but negatively correlated with Acanthamoeba spp. and M. intracellulare. P. aeruginosa concentrations were positively correlated with increased time to the last rain event (τ = 0.142), but negatively correlated with the last rainfall event depth (τ = -
0.176) and total rainfall during the week prior to the sampling event (τ = -0.180).
4.6. Discussion
OPPPs are key constituents of engineered water systems due to their epidemiologic importance, widespread occurrence in waterborne environments, resistance to disinfection, and complex microbial ecology (Pruden et al. 2013). The first phase of this study previously identified high concentrations of six OPPPs in RHRW tanks in Queensland (Hamilton et al.
2016). For the twenty-four tanks chosen for six monthly follow-up events, most sampling events were taken within 48 h of a rainfall event and all were taken within 7 days of a rain event with the exception of nine samples in February. The February sampling time point was assessed in order to maintain the seasonal sampling scheme every 3-4 weeks. Although rainfall has been shown to correlate with fecal pollution in rainwater tanks (Vialle et al.
2011), OPPPs are non-fecal, environmental (saprozoic) microorganisms and rather are influenced by localized and ambient growth conditions including system materials, configuration and hydraulic conditions, disinfectant concentration, temperature, and nutrient availability, and microbial ecology (Falkinham 3rd et al. 2015, Wang et al. 2012a). Therefore, the emphasis of the sampling scheme was to study the seasonal occurrence of OPPPs in tank water samples.
The prevalence of FIB and OPPPs indicate that with the exception of L. pneumophila, a clear pattern of recurring contamination of specific tanks was not present over a six-month period. For FIB, this could be due to episodic fecal pollution associated with the sporadic presence of wildlife on roofs or rain events, and rapid (3-4 days) decay periods that could allow FIB to reach concentrations below the method limit of detection by the time of sampling(Ahmed et al. 2014c). For OPPPs, L. pneumophila was present in the same tanks over time when it was detected. These data support that conditions in specific tanks were
73
conducive to the growth of L. pneumophila or its relative success in a particular ecologic niche within these tanks. The tanks where L. pneumophila was present (T8, T22, T24) were two concrete and one steel tank present on their respective properties for >5 years, infrequently cleaned, and receiving direct sunlight for more than half of the day. All three locations had clay/concrete tile roofs. These factors may have contributed to their presence in tank water samples. However, due to the small number of detection events for L. pneumophila
(n = 7), limited comparisons can be made.
Some seasonal differences between the prevalence of certain FIB and OPPPs during the wet and dry seasons (Wells 2013) were observed, with more samples positive for FIB and higher average FIB concentrations during the wet season months. The presence and concentration of FIB were positively correlated with higher ambient temperatures and rainfalls, and for Enterococcus spp., concentrations were also positively correlated with higher RH. A previous study by Vialle et al. (Vialle et al. 2011) measured FIB in a single tank in rural south-western France from January 2009 to January 2010, noting a correlation between E. coli and Enterococcus spp. with daily rainfall and supporting rainfall as the mechanism for introducing FIB to tanks. Higher rainfall, RH, and ambient temperatures are also more likely to occur during the wet season. Although increased temperatures may increase inactivation of FIB on roof surfaces, once introduced to tanks, decay could be much slower than on the roof and allow the FIB to persist (Ahmed et al. 2014c).
OPPPs varied in their seasonal occurrence. Legionella spp. and M. intracellulare were consistently present across sampling events, however both OPPPs occurred at higher concentrations during the dry season (Sep-Nov for Legionella spp. and Aug-Sep for M. intracellulare). Acanthamoeba spp. and M. avium occurred sporadically but peaked in prevalence in September and October, respectively. P. aeruginosa prevalence and concentrations peaked during the wet season. In particular, P. aeruginosa was the only microorganism that was positively and significantly correlated to the time since the last rainfall event. Stagnation time affects OPPP growth and detection(Wang et al. 2012b) and could potentially play a greater role in enhancing the occurrence of P. aeruginosa compared
74
to other OPPPs observed in this study. L. pneumophila was only detected during the wet season. This could be due to slightly elevated RH observed during the wet season sampling events. RH on the sampling day was weakly correlated with increased concentrations of L. pneumophila in this study. This may be concerning since the occurrence of cases of
Legionnaires’ Disease also correlates with high humidity (Fisman 2005).
Consistent with the findings of FIB and OPPPs seasonality, FIB were significantly correlated with P. aeruginosa, but not other OPPPs. This indicates that although some OPPPs such as P. aeruginosa may possess growth requirements or other ecological preferences similar to FIB, FIB are generally not good indicators of the presence of OPPPs. This is consistent with previous findings in RHRW (Ahmed et al. 2014b, Hamilton et al. 2016).
Furthermore, a previous study of Legionella spp. in swimming pool showers noted an inverse correlation with P. aeruginosa and also other gram negative bacteria at temperatures < 43ºC
(Leoni et al. 2001). Positive correlations were observed between Legionella spp. and M. intracellulare with Acanthamoeba spp., which is consistent with known symbiotic relationships between multiple species of the Legionella genus as well as MAC with protozoans such as Acanthamoeba that encourage bacterial proliferation, persistence, and virulence (Buse and Ashbolt 2012, Cirillo et al. 1997, Kuiper et al. 2004a, Thomas et al.
2010). Positive correlations between MAC and Legionella spp. may be indicative of the similar ecologic niches occupied by these two groups of bacteria (Pryor et al. 2004a); similar correlations have been noted in reclaimed water (Jjemba et al. 2010) and drinking water distribution systems (Wang et al. 2012a).
The human health significance of the high concentrations of OPPPs observed depends on the type and frequency of exposure, and is typically determined using quantitative microbial risk assessment (QMRA). The concentrations of L. pneumophila observed in the current study
(2.3 x 102- 1.5 x 103 gene copies per L) would be lower or comparable to concentrations associated with a commonly cited risk benchmark of 1 extra infections per 10,000 people per year for shower water (3.5 × 106 to 3.5 × 108 per L) (Schoen and Ashbolt 2011) and 6-170 gene copies per L in water used for showering or hosing (Ahmed et al. 2010). However, the
75
observed concentrations are comparable to those associated with excess risks for 2 h exposures in recreational splash parks using rainwater as a source (1,200 CFU per L) and whirlpool spas (10 CFU per L) (Armstrong and Haas 2007b, 2008, Bouwknegt et al. 2013, de
Man et al. 2014a). This could be problematic if rainwater is used for pool top-up as reported in the survey. It is noted here that L. pneumophila is not the only species of Legionella that can cause disease, and over 19 species have been documented as human pathogens (Muder and Victor 2002). This warrants further exploration of Legionella diversity and non- pneumophila species that might play a role in RHRW-associated disease, as observed
Legionella spp. concentrations were high (3.2 × 103 to 2.3 × 107 gene copies per L).
Concentrations of P. aeruginosa were above a common benchmark density for swimming pool corrective action of 1 CFU/ 100 mL (World Health Organization 2006). The currently observed concentrations of P. aeruginosa (3.6 × 102 to 4.7 × 108 gene copies per 100 mL) are also comparable to those associated with skin disease outbreaks (Roser et al. 2015).
Concentrations of MAC associated with hypersensitivity outbreaks in hot tubs have ranged from 1.4 × 102 to 4.3 × 104 CFU per 100 mL (Fjällbrant et al. 2013, Lumb et al. 2004), which is comparable to M. avium (2.4 × 101 to 3.6 × 105 gene copies per 100 mL) and M. intracellulare (2.2 × 101 to 6.8 × 105 gene copies per 100 mL) concentrations observed in the current study, warranting further examination of OPPP risks.
Limited studies on longitudinal or seasonal occurrence of FIB in tanks are available, and no study to date has examined OPPP occurrence in multiple tanks over time. Only Vialle et al. (Vialle et al. 2012, Vialle et al. 2011) measured FIB and opportunistic and enteric pathogens using culture-based methods in a single tank from rural south-western France weekly over a one-year period, detecting L. pneumophila once at 700 CFU per L, and Giardia once (1 cyst per 20 L). Half of the samples contained Aeromonas spp. (43%, n = 28) and P. aeruginosa (41%, n = 17), but the temporal variations of the samples was not mentioned. The lower occurrence of pathogens in some of these studies may be due to the use of culture- based methods that are known to underestimate pathogen concentrations due to the formation of viable but non-culturable (VBNC) cells (Moritz et al. 2010, Oliver 2005). For this reason,
76
the qPCR methods used in this study offer advantages for quantifying OPPPs in tank water samples. However, they do not provide information on viability and/or infectivity of OPPPs, which remains a limitation.
The seasonality of OPPP occurrence is valuable for informing rainwater tank owners, as management options are limited. Gutter guards may be implemented to limit plant material (a carbon source) from entering tanks and increasing microbial growth, however the impact in terms of the impact on pathogen concentrations has not been investigated. A first-flush device is another option that diverts the first volume of rainwater to avoid reaching the tank, as this water contains the greatest concentration of contaminants (Bertrand-Krajewski et al. 1998).
However, the recommended volume of water to be removed and type of recommended removal device can vary depending on the rainwater harvesting system characteristics
(enHealth 2004, Lee and Visscher 1992, TWDB 2005, WHO 2004). Although first flush devices have shown some ability to remove microorganisms and other contaminants (Amin and Han 2011, Doyle 2008, Lee et al. 2012, Yaziz et al. 1989), they have not been shown to be completely effective for preventing contamination of tanks as there is incomplete wash-off during the initial stage of relatively smaller intensity rainfall events (Egodawatta et al. 2009,
Kus et al. 2010, Mendez et al. 2011). In one case, the removal efficacy of first flush devices was been shown to be satisfactory for physicochemical quality but not microbiological quality
(Gikas and Tsihrintzis 2012). Unlike physicochemical contaminants, microorganisms can also accumulate in sediments and biofilms, and regrow under the warm conditions likely to be present in Southeast Queensland. Additionally, if the primary reason for the appearance of opportunistic pathogens is proliferation within the tank itself, the reduction of inoculum will only serve to increase the lag time to proliferation rather than reducing proliferation itself.
Additional research is needed to assess which mechanism (initial inoculation of tank by pathogen sources or growth/decay within the tank) predominates. As a result, risk management strategies for FIB and OPPPs may differ; while increased water age can foster decay of FIB, it can also encourage the proliferation of OPPPs(Wang et al. 2012b). For this reason, it is valuable to identify characteristics associated with the attenuation of both types of
77
microorganisms. However, none of the meteorological characteristics observed in this study were consistently negatively associated with the occurrence and concentrations of both FIB and OPPPs.
Apart from maintenance of the catchment area, tank, and gutters, point-of-use (POU) devices are one of the only other remaining water quality management options. However,
POU for rainwater may also not be sufficient for bringing water in compliance with drinking water guidelines, especially for pathogens (Dobrowsky et al. 2015a, Dobrowsky et al. 2015b,
Jordan 2008, Reyneke et al. 2016). Additionally, limited studies have assessed inactivation of
FIB and enteric pathogens in tanks under ambient environmental and system temperature conditions (Ahmed et al. 2014c, Spinks et al. 2006), but no studies have addressed OPPP removal or regrowth after treatment application. As the control of OPPPs may be targeted differently than for enteric pathogens (Wang et al. 2013), this warrants further examination using both qPCR and culture-based methods for mitigation of potential RHRW health risks.
78
5. Dose response models for Mycobacterium avium complex (MAC)3
5.5. Abstract:
Mycobacterium avium complex (MAC) is a group of environmentally-transmitted pathogens of great public health importance. This group is known to be harbored, amplified, and selected for more human-virulent characteristics by amoeba species in aquatic biofilms. However, a quantitative microbial risk assessment (QMRA) has not been performed due to the lack of dose response models resulting from significant heterogeneity within even a single species or subspecies of MAC, as well as the range of human susceptibilities to mycobacterial disease. The primary human-relevant species and subspecies responsible for the majority of the human disease burden and present in drinking water, biofilms, and soil are M. avium subsp. hominissuis, M. intracellulare, and M. chimaera. A critical review of the published literature identified important health endpoints, exposure routes, and susceptible populations for
MAC risk assessment. In addition, data sets for quantitative dose-response functions were extracted from published in vivo animal dosing experiments. As a result, seven new exponential dose response models for human-relevant species of MAC with endpoints of lung lesions, death, disseminated infection, liver infection, and lymph node lesions are proposed. Although current physical and biochemical tests used in clinical settings do not differentiate between M. avium and M. intracellulare, differentiating between environmental species and subspecies of the MAC can aid in the assessment of health risks and control of MAC sources. A framework is proposed
3 This chapter is published: Hamilton, K.A., Weir, M.H. and Haas, C.N. (2017) Dose response models and a quantitative microbial risk assessment framework for the Mycobacterium avium complex that account for recent developments in molecular biology, taxonomy, and epidemiology. Water Research 109, 310-326. 79 for incorporating the proposed dose response models into susceptible population- and exposure route- specific QMRA models.
5.6. Keywords:
Mycobacterium avium complex, Mycobacterium intracellulare, Mycobacterium chimaera, MAC, Quantitative Microbial Risk Assessment, QMRA, dose response, opportunistic pathogen, biofilm, engineered water system, aerosol
5.7. Introduction
In recent years, improvements in disinfection practices have reduced the health burden of diarrheal pathogens in drinking water systems in many developed and some developing countries. As a result, the focus for waterborne disease burden mitigation in these regions has shifted to opportunistic pathogens that live in biofilms growing on the inner surfaces of distribution system and premise plumbing pipes such as Legionella spp., Mycobacterium spp., and Pseudmonas aeruginosa, among others
(Falkinham 3rd et al. 2015, Pruden et al. 2013). For example, Legionella spp., a known inhabitant of these systems, now represents the most common cause of drinking water outbreaks in the United States (Beer et al. 2015). Opportunistic premise plumbing pathogen-related illness is likely to represent a lesser portion of the waterborne disease burden in developing countries compared to developed ones, however, outbreaks and isolations from environmental media in developing countries have been reported (Bartram et al. 2007, Pavlik et al. 2009c, von Reyn et al. 1993b).
While significant attention has been devoted to the study of health risks from exposure to Legionella spp. in engineered water systems (Armstrong and Haas 2007a,
Armstrong and Haas 2007b, Schoen and Ashbolt 2011), a framework has not yet been
80 developed for quantifying risks due to Mycobacterium spp., an increasingly important cause of opportunistic infections (Falkinham et al. 2015, Pavlik et al. 2009d).
The genus Mycobacteria contains over 150 species (Tortoli 2003) and is divided into human/animal obligate pathogens including tuberculosis-causing mycobacteria, and non-tuberculosis mycobacteria (NTM4) (Portaels 1995, Vaerewijck et al. 2005). Most mycobacteria are 2 – 5 µm long and 0.2 – 2.0 µ thick (Pavlik et al.
2009c). Mycobacterium avium complex (MAC) is a group of related species of non- tuberculosis mycobacteria (Portaels 1995, Vaerewijck et al. 2005) listed on the United
States Environmental Protection Agency contaminant candidate list (CCL3 and Draft
CCL4) (USEPA 2009, 2015). MAC is the most frequently identified cause of the 4.2 -
7.2 per 100,000 annual pulmonary infections due to NTM in the United States in the general population and 15 - 47 per 100,000 in elderly (> 65 years of age) populations, but is not a United States Centers for Disease Control and Prevention (CDC) reportable illness (Adjemian et al. 2012b, Cassidy et al. 2009, CDC 2011,
Kasperbauer and Daley 2008, O'Brien et al. 1987, Winthrop et al. 2011).
Additionally, MAC has been associated with several hospital acquired infections and healthcare outbreaks (Aronson et al. 1999, Tobin-D’Angelo et al. 2004b). It is environmentally transmitted, and person-to-person transmission is not believed to occur. In addition to pulmonary disease, various disease outcomes are associated with
MAC including soft tissue infections and cervical lymphadenitis in immune- competent patients, and disseminated infections in immunocompromised patients
(Falkinham III 1996). MAC pulmonary disease is rare in children and usually related to other immune deficiencies, however, MAC cervical lymphadenitis is the most
4 NTM is often also referred to as atypical mycobacteria, environmental mycobacteria, potentially pathogenic environmental mycobacteria, or mycobacteria other than tuberculosis
81 common form of NTM disease in children (Lai et al. 1984, Lincoln and Gilbert 1972).
Few epidemiologic studies of NTM in children have been conducted, but those from developed countries indicate an annual incidence ranging up to 5.7 NTM infections per 100,000 children under 5 years from Sweden (Lopez-Varela et al. 2015, Romanus et al. 1995).
The public health importance of MAC is increasing; it is suggested that the disease prevalence and laboratory isolation is increasing even after considering its evolving taxonomy, increased awareness, improved laboratory analytical methods
(Appendix Table 9.16), and improved clinical diagnostic tools (Johnson and Odell
2014, Kasperbauer and Daley 2008, Khan et al. 2007). Some of this increase may be due to recognition of the role of MAC in cystic fibrosis and bronchiectasis, increased use of showers compared to baths, and ageing of the population (Angrill et al. 2001,
Barker 2002, Field et al. 2004, Kilby et al. 1992, Oliver et al. 2001, Prince et al.
1989). In addition, increased prevalence of diseases which decrease immunocompetence (HIV/AIDS and cancer), increased use of chemotherapeutic drugs for cancer treatment resulting in immunosuppression, lifestyle changes that bring humans into contact with habitats where NTM naturally occurs, improved water treatment that plays a role in selecting for pathogenic NTM, and changes to the climate and environment may also play a role (Pavlik et al. 2009b). MAC is widespread in waterborne environments and soil (especially those containing peat or sphagnum vegetation) (Falkinham III et al. 2001, Rusin et al. 1997, Tuffley 1980,
Vaerewijck et al. 2005, von Reyn et al. 1993b), as well as animal-derived foods
(Klanicova et al. 2011) and produce (Cerna-Cortes et al. 2015). It is associated with engineered water systems and biofilms, contributing to its resistance to disinfectants and behavior as an intracellular parasite of free-living protozoans such as
82
Acanthamoeba spp. in a similar manner to Legionella spp. (Berry et al. 2010, Cirillo et al. 1997, Drancourt 2014, Falkinham 2013, Steinert et al. 1998, Wang et al. 2012a,
Whiley et al. 2012).
The number of species classified as MAC are increasing with advances in genetic sequencing, but currently include M. avium, M. intracellulare, M. arosiense,
M. chimaera, M. colombiense, M. marseillense, M. timonense, M. bouchedurhonense, and M. ituriense (Falkinham 2013). M. avium consists of four subspecies
(paratuberculosis (MAP), avium, hominissuis, and silvaticum) (Vaerewijck et al.
2005). MAC has 28 serotypes (Wolinsky and Schaefer 1973) comprised of M. avium susp. avium (1-3), M. avium subsp. hominissuis (4-6, 8-11, 21), and M. intracellulare
(7, 12-20, 22-28). Not all MAC species and subspecies are significant causes of human disease. The primary human-relevant species and subspecies responsible for the majority of the human disease burden and present in drinking water, biofilms, and soil are M. avium subsp. hominissuis, M. intracellulare, and M. chimaera (Falkinham
III 2013a, Hibiya et al. 2011, Rindi and Garzelli 2014, Wallace et al. 2013) (Table
5.1). M. avium subsp. paratuberculosis has additionally been isolated from drinking water (Beumer et al. 2010), and may also be relevant for water and wastewater practitioners as it has been implicated as a potential causative agent of Crohn’s disease in humans, although this relationship is contentious (Pierce 2009, Waddell et al. 2015).
83
Table 5.1 MAC primary host organisms and disease outcomes (adapted from Hibiya et al. (2011) and Rindi and Garzelli (2014) unless otherwise noted)
Species/ subspecies Major Host Principal environmental Disease outcomes reservoir M. avium M. avium subsp. Ruminants Johne’s Disease; potential links to Crohn’s disease in paratuberculosis humans M. avium subsp. avium Birds Avian TB M avium subsp. Wood pigeons TB-like disease silvaticum M. avium subsp. Humans, pigs Drinking water, biofilms Various clinical outcomes depending on host immunity; hominissuis pulmonary infections, cervical lymphadenitis, soft tissue infections, disseminated infections M. intracellulare Humans, ruminants, Soil (De Groote et al. 2006) Respiratory and pulmonary infections pigs M. arosiense Human Osteomyelitis, pulmonary disease M. chimaera Human Drinking water, biofilms Pulmonary infections M. colombiense Human Disseminated infections, lymphadenopathy M. marseillense Human Pulmonary infections M. timonense Human Pulmonary infections M. bouchedurhonense Human Pulmonary infections M. ituriense Unknown Unknown (Falkinham 2013, Salah et al. 2009) M. vulneris Human Lymphadenopathy, wound infections “MAC-others” Human Pulmonary infections, disseminated infections 84
Making policy and engineering decisions to mitigate pathogen risks based upon epidemiologic data is challenging due to limitations in detecting small changes in health outcomes, and high costs associated with studies large enough to do so.
Additionally, infections with long incubation periods such as MAC are often difficult to attribute to a source (Pavlik et al. 2009b). As a result, quantitative microbial risk assessments (QMRA) are increasingly used in order to inform public health regulatory decisions for water- and food-borne pathogens (Ashbolt et al. 2010,
Hathaway and Cook 1997). This approach utilizes dose-response modeling to infer the effects of (usually low-dose) exposures to pathogens using a four-step process of hazard identification, exposure assessment, dose response, and risk characterization
(Haas et al. 1999). To develop appropriate public health management strategies,
QMRA researchers, risk assessors, and environmental regulators require quantitative dose-response relationships to determine the potential for pathogens to cause harm given their occurrence under a set of environmental and human exposure conditions.
Dose response models for the most human-relevant strains of MAC have not previously been developed in part due to significant heterogeneity within even a single species or subspecies of MAC (Falkinham III 2013b) and significant variation in human susceptibilities to mycobacterial disease (Casanova and Abel 2002).
Infectivity studies and development of dose response models for MAC have been identified as a top priority for opportunistic pathogens in premise plumbing by the
American Water Works Association, the Water Research Foundation, and other researchers (AWWA 1999, Falkinham 3rd et al. 2015, Khan et al. 2007, Pruden et al.
2013). Development of such models that explore variability in strain virulence will inform future infectious dose studies of MAC and fill a critical research gap necessary to complete a QMRA. Filling this key research gap will allow a QMRA to be 85 performed, which in turn will answer questions about how MAC risk can be best mitigated and/or managed. A dose response model is available for a single subspecies of MAC, M. avium paratuberculosis (Breuninger and Weir 2015), but in being for only one subspecies cannot be used to model risks from the complete complex.
The purpose of this research is therefore to develop new dose response models for the most epidemiologically-important non-MAP subsets of MAC with known linkages to engineered water systems and soils (Boyle et al. 2015, Donohue et al.
2015, Lu et al. 2016, Thomson et al. 2013b): M. avium subsp. hominissuis, M. intracellulare, and M. chimaera. To do so, we perform a critical review of the published literature to identify important health endpoints and data sets for quantitative dose-response functions. A resulting framework for MAC QMRA is postulated, and research gaps for both MAC dose response function development and
QMRA are identified.
5.8. Literature review
5.8.6. Search strategy
To compare infectivity among MAC species and subspecies, a literature review was conducted to identify datasets for M. avium subsp. homissuis, M. intracellulare, or M. chimaera generated from MAC animal experiments. Google
Scholar, Web of Science, Agricola, PubMed, Engineering Village, and Medline(Ovid) were searched using search terms: “((Mycobacterium avium) OR (M. avium) OR
(Mycobacterium intracellulare) OR (M. intracellulare) OR (Mycobacterium chimaera) OR (M. chimaera) OR (Avian tubercle bacillus) OR (M. tuberculosis avium) OR (Battey bacillus) OR (Nocardia intracellularis) OR (Mycobacterium battey) OR (M. avium-intracellulare) OR (M. avium-intracellulare-scrofulaceum))
86
AND ((susceptibility) OR (pathogenesis) OR (challenge model) OR (volunteer challenge) OR (in vivo) OR (feeding) OR (experimental challenge) OR (infectious dose) OR (inoculation) OR (aerosol infection) OR (experimental mycobacteriosis)
OR (experimental illness) OR (fecal shedding) OR (volunteer response) OR
(bacteriologic response) OR (mouse model) OR (mice response) OR (porcine model)
OR (pig model) OR (inoculum response)) NOT ((paratuberculosis) OR (johne disease) OR (johne’s disease) OR (Johne* disease))”. Search results are summarized in Appendix Table 9.17.
Avian tubercle bacillus, M. tuberculosis avium, Battey bacillus, Nocardia intracellularis, M. battey, Battey-avian mycobacteria, M. avium-intracellulare, and
M. avium-intracellulare-scrofulaceum are historical names for the general MAC, which were therefore included in this search based on Turenne et al. (2007). Relevant records were forward and reverse citation searched and imported into an EndNote® library with duplicate removal. 2,842 records were obtained and reviewed for the inclusion criteria below. 397 records for which full text could be obtained described in vivo models which were the focus of this preliminary work. Inclusion criteria were as follows:
Number of organisms in the dosing inoculum is quantified;
Criteria for a positive endpoint is stated and monitored for;
Number of subjects administered the inoculum and experiencing the
endpoint (illness, infection, death) is defined and quantal5;
Dosing method and exposure route defined (intravenous, oral,
intratracheal, etc.);
5 Quantal responses are binary, therefore the number of animals exhibiting a given endpoint (illness, death, etc.) in each dose group is yes/no (1 of 4 died, for example).
87
Strain and/or source of inoculum is defined (deer, chicken, human, etc.)
and determined to be of human relevance (belonging to M. avium subsp.
hominissuis or M. avium obtained from a human-derived source, M.
intracellulare, or M. chimaera; see (Saito et al. 1990) and (Wayne et al.
1993) for historical strain descriptions);
One or more intermediate responses are observed (not 0% or 100%
infected).
5.8.7. Curve fitting
All datasets meeting initial inclusion criteria are summarized in Appendix Table 9.18.
A Cochran-Armitage test of trend (Weir and Haas 2009) was performed for individual and pooled datasets to determine if further model fitting was appropriate via maximum likelihood estimation and bootstrapping with 10,000 iterations using the exponential (Equation 5.1) and Beta-Poisson (Equation 5.2) dose response models
(Haas et al. 1999).
푃(푟푒푠푝표푛푠푒) = 1 − 푒−푟푑 Equation 5.1
−훼 21/훼−1 푃(푟푒푠푝표푛푠푒) = 1 − [1 + 푑 ( )] Equation 5.2 푁50
Where r= probability of an organism surviving and reaching the appropriate site to initiate infection; d= organism dose; α, β, N50 are parameters of the Beta-Poisson model and the N50 is the dose at which 50% of the population is expected to be affected. The optimization uses the maximum likelihood estimation (MLE) method to minimize the negative-2 log-likelihood (deviance) of the model. If the deviance was
88
less than the 95% confidence value for the 2 distribution with degrees of freedom equal to the number of doses minus the number of parameters in the model, then the models were considered a good fit to the data set. The null hypothesis of fit acceptability is therefore rejected (i.e. the dose-response model is rejected) for P <
0.05 (Haas et al. 1999). The best-fitting model was identified by the model that produces a difference in the deviances of the two models (Δ) compared to the 95% confidence interval for the 2 distribution at one degree of freedom.
In some cases, pooling of datasets is possible where two or more datasets have the same observed clinical endpoint (lung infection, for example). To evaluate if the pooled model provided a better fit compared to individual M. avium or M. intracellulare datasets, the difference in deviances, ∆푌, was assessed as per equation
(5.3).
푛 ∆푌 = 푌푇 − ∑푖 푌푖 Equation 5.3
where, 푌푇 is the fit of the common model to the pooled data and 푌푖 are the fits of the individual models. To determine if there is a benefit to the fit of the model due to pooling the data the ∆푌 is compared to the χ2 distribution at the same confidence with m degrees of freedom (equation 5.4). If the ∆푌 is greater than the χ2 critical value then there is a statistically significant improvement in fit from pooling the data. The optimizations and subsequent analyses were performed in the R software environment
(www.rproject.org).
푛 ∆푚 = ∑푖 푚푖 − 푚푇 Equation 5.4
89
where, 푚푇 is the number of parameters in the pooled dose-response fit and 푚푖 are the number of parameters in the individual dose-response fits.
5.9. A framework for MAC quantitative microbial risk assessment
5.9.6. MAC exposure routes
Aerosolized water is likely to be the most significant source of human MAC infection (Field et al. 2004, Goslee and Wolinsky 1976) and has been documented using matching of clinical and water samples (von Reyn et al. 1994). Falkinham III
(2013a) recently identified additional possible routes of NTM infection (including
MAC) as: 1) ingestion of soil or water by children leading to cervical lymphadenitis,
2) inhalation of NTM-laden aerosols by adults leading to pulmonary infection, 3) inhalation of dusts from potting soils leading to pulmonary infection, 4) aspiration due to gastric reflux of NTM entering the stomach via ingestion leading to pulmonary infection, and 5) oral ingestion of NTM in water by patients with profound immunodeficiency and reduction of the normal barrier function of the gastrointestinal tract leading to disseminated infection. Additionally, iatrogenic routes of MAC have been documented via bronchoscopy, catheters, transplants, and therapeutic pool, hot tub, and nebulizer users (Gubler et al. 1992, Sax et al. 2015). M. avium and M. chimaera are likely to be transmitted by water, while M. intracellulare is likely to be transmitted through soil (Wallace et al. 2013). The most likely exposure route in non-
AIDS patients is therefore the respiratory tract (resulting in non-disseminated infection), while in AIDS patients it is the gastrointestinal tract, resulting in disseminated infection (Bermudez 1994). A list of studies of environmental media
90 from which relevant MAC species have been identified is provided in Appendix Table
9.19.
5.9.7. MAC pathogenesis
The most common clinical manifestation of NTM is pulmonary disease
(Donohue et al. 2015). Despite the importance of environmental exposure route, host characteristics may be relatively more important factors in MAC lung disease pathogenesis. A case-control study of 52 matched pairs reported that aerosol- generating activities and home water features were not associated with disease, while prior lung disease and immune-suppressing drugs were associated with susceptibility to MAC lung disease (Dirac et al. 2012). Historically, MAC was most frequently reported via single-strain, infection of older males with pre-existing lung disease conditions, “dusty” occupations, and tobacco or alcohol use and clinical presentation via bilateral upper lobe fibrocavity disease (Field et al. 2004). However, MAC infection has recently been reported more commonly in healthy older females, whom may be tall, thin, and post-menopausal, with an outcome of fibronodular disease with bronchiectasis, also known as “Lady Windermere’s Syndrome” (Chan and Iseman
2010, Field et al. 2004, Kartalija et al. 2013, Kim et al. 2008, Mirsaeidi and Sadikot
2015). This epidemiologic transition has been reported to have occurred from the
1940’s to 1980’s (Prince et al. 1989, Young and Inderlied 1990) and continuing through the 1990’s (Falkinham III 2016, Wolinsky 1979), but the exact reason for this transition has not been confirmed. Finally, healthy, young patients with no predisposing conditions exposed to MAC may develop hypersensitivity pneumonitis or “hot tub lung”, often exhibiting coinfection with multiple strains of MAC (Embil et al. 1997, Field et al. 2004, Fujita et al. 2014, Mangione et al. 2001, Rickman et al.
91
2002, Sugita et al. 2000). However, respiratory illness in this case is thought to be due to hypersensitivity rather than infection (Embil et al. 1997, Marras et al. 2005).
5.9.8. Predominating MAC species
Multiple species/subspecies/clones of MAC have been isolated from patients; mixed (coinfection with multiple species) or polyclonal infection (varied strains of the same species) are common, and the specific MAC species causing the infection is not typically distinguishable based on the disease pathology alone (Arbeit et al. 1993,
Falkinham III 1996). Testing is also not typically performed due to the large number of potential NTM species, lack of standardized testing approaches, and absence of clinical advantage. However, recent efforts using multilocus sequencing analysis support the notion of varying degrees of virulence between species and distinct disease outcomes (Boyle et al. 2015). Patients with polyclonal infections have been associated with higher soil exposure, more frequent use of bathroom showers, and more frequent swimming in pools than those with monoclonal infections (Fujita et al.
2014). During coinfection, the predominating strain of MAC further varies depending on host immune status; for example M. avium predominates over M. intracellulare in
AIDS vs. non-AIDS patients (Crowle et al. 1992, Guthertz et al. 1989). M. avium is likely to predominate over other strains in cases of cervical lymphadenitis (Hazra et al. 2000, Kyriakopoulos et al. 1997, Oloya et al. 2008). However, this pattern has changed over time as M. scrofulaceum was previously the most common cause of lymphadenitis in children (Wolinsky 1979); absence of M. scrofulaceum in natural waters from the southeastern United States which previously displayed high concentrations (von Reyn et al. 1993b) indicates a change in the geographic distribution of mycobacterial strains (Falkinham III 1998). Additionally, the transition to M. avium from M. scrofulaceum is thought to be due to increased chlorine
92 disinfection rates in drinking water, which may have contributed to changes in the mycobacterial species profile of drinking water (Primm et al. 2004). For older females with nodular bronchiectasis, M. intracellulare is the most commonly isolated species
(Han et al. 2005, Wallace Jr et al. 1998). Furthermore, patterns of dominant MAC strains vary temporally, geographically, and seasonally (Adjemian et al. 2012a,
Brooks et al. 1984a, Falkinham III et al. 1980, Inderlied et al. 1993, Ito et al. 2015,
Kirschner Jr et al. 1992, Martin-Casabona et al. 2004, O'Brien et al. 1987, Shao et al.
2015, Thomson et al. 2013b, Tsang et al. 1992). For example, a recent study demonstrated that in a group of 448 Chicago patients who provided MAC-positive pulmonary specimens from 2000-2012, the proportion of samples positive for M. chimaera generally increased over this time comparative to M. avium and M. intracellulare (Boyle et al. 2015).
Differing information is available regarding the predominant strain in cases of cavitary disease. In the Chicago study, 15% of total patients were diagnosed with cavitary disease (Boyle et al. 2015). Of the patients who tested positive for M. avium
(n=241) and who had radiographic testing performed (n=236), 14% had cavitary disease compared to 13% of M. intracellulare-positive patients (81 total positive patients, radiographic testing performed on 80 patients). Twenty-percent of M. chimaera- positive patients had cavitary disease. Patients with pulmonary infection with M. chimaera were more likely to be immunosuppressed than those infected with other MAC strains (Boyle et al. 2015). Another study from Texas identified M. intracellulare as the dominant strain in nine cavitary disease patients (Wallace Jr et al.
1998). However, a Japanese study compared 10 patients with M. avium lung disease to 36 patients with M. intracellulare lung disease, finding that in the M. avium patients, cavitary lung disease was more common (Tanaka et al. 2000). Another
93
Japanese study noted that 82% (n=29) and 72% (n=43) patients with M. avium and M. intracellulare lung disease had chest radiographic findings of cavitary disease, respectively (Maesaki et al. 1993). Finally, a South Korean study noted higher prevalence of the nodular bronchiectasis form of MAC disease compared to other forms in both 323 M. avium and 267 M. intracellulare positive patients with lung disease (Koh et al. 2012). Acknowledging the complexities summarized above, a framework for QMRA of MAC is presented in Figure 5.1. These pathways can be used to select one of the proposed dose response models depending on the susceptible population and disease pathway considered.
94
aMost significant exposure routes shown here. Exposure is also possible through injuries to the skin although this is less common and is therefore not included in the diagram. bMost significant populations at risk shown here. Note that other populations with underlying lung disease and various susceptibilities apart from the groups shown here are also vulnerable to MAC infection.
Figure 5.1 Mycobacterium avium complex (MAC) quantitative microbial risk assessment (QMRA) framework 95
5.10. Dose response models for MAC
5.10.6. Animal models
Clinical dosing studies of pathogens in humans are considered the “gold standard” for dose-response relationships. However, it would not be feasible to conduct these types of studies with patients vulnerable to opportunistic infections. Therefore, animal models of infection can be used to provide this information, and are frequently used within the dose response methodology of the QMRA paradigm (Haas et al. 1999).
Of the 396 in vivo whole animal studies obtained from literature review, results from those meeting the described inclusion criteria and providing acceptable exponential or beta-poisson dose response model fits are summarized in Table 5.2.
Model fits are summarized in Table 5.3 and presented in Figure 5.2-5.8. Pooling across the lesions endpoint for Tomioka et al. (1993b) and Jorgensen (1977) did not provide an improvement in fit compared to the individual models. Although some in vitro models exist for MAC hypersensitivity (Huttunen et al. 2001, Huttunen et al.
2000), whole animal models of hypersensitivity pneumonitis after exposure to MAC were not found during the literature review searches. 96
Table 5.2. Data used in the current study for MAC dose response modeling
Source Host/Age Route Endpoint Strain Time to Dose (CFU) No. Hosts Positive Responses Negative Responses followup (Tomioka et al. Female Intravenous Lung lesions M. avium 363 days- 1.3 × 106 6 3 3 1993a) C57BL/6 serovar 1, serovar 1 2.1 × 106 6 6 0 beige Mice/ 5 human non- 2.2 × 106 6 6 0 weeks AIDS isolates: 3.0 × 106 6 6 0 pooled strains 4.7 × 106 6 6 0 N-289, N-356, N-357, N-364, 363 days- 0.7 × 106 6 2 4 N-445, N458, serovar 9 2.9 × 106 6 6 0 N-461; Serovar 9 strains N- 254, N-302 (Yangco et al. Male golden Intratracheal Disseminated MAC serotype 24 weeks 1× 107 59 21 38 1989) Syrian inoculation infection 8 (Human- 1× 108 59 37 22 hamsters/ 12 derived, AIDS 5× 108 60 41 19 weeks patient)
Liver infection 24 weeks 1× 107 59 5 54 1× 108 59 8 51 5× 108 60 15 45 (Mehta 1996) C57BL/6 Intravenous Death MAC 101 30 days 1× 106 20 0 20 beige mice/ (human- 1× 107 20 0 20 NS derived) 2× 107 20 0 20 5× 107 20 4 16 1× 108 20 6 14 (Jorgensen 1977) Pigs/ 8-10 Intravenous Lesions or culture M. avium 6-8 months weeks old in lymph nodesa serotype II 1 strain SSC 7.8× 104 2 0 2 1323(porcine- 7.8× 105 2 0 2 derived) 7.8× 106 2 0 2 7.8× 107 2 1 1 3.9× 108 2 2 0
2 7.8× 104 2 0 2 7.8× 105 2 0 2 7.8× 106 2 1 1 7.8× 107 2 2 0 3.9× 108 2 2 0
3 7.8× 104 2 0 2 7.8× 105 2 0 2 7.8× 106 2 1 1 7.8× 107 2 1 1 3.9× 108 2 2 0 97
a1. Lesions in lymph nodes: mandibular, parotideus, retropharyng. lat. retropharyng. med., cervicalis sup. dors., cervicalis sup. ventr., subiliacus, popliteus, inguinalis prof., inguinalis superf., spleen, tonsil. Culture from lymph nodes: cervicalis sup., cervicalis sup. ventr., subiliacus, ingninalis prof., igninalis superf. liver, kidney; 2. Lesions in lymph nodes: Liver, Intestinal mucosa (Peyer Patch); 3. Culture from lymph nodes: Retropharyng. lat.; NS: Not specified in the original reference
98
Table 5.3. Model fit results for data in Table 5.2 (exponential model parameter r, Beta-Poisson model parameters α, N50).
a c Model Dose Gps Parameter values Minimized df= #dose Χcrit Acceptible fit? Preferred model p, fit Deviance (Y) groups- # param Tomioka et al. 7 r=1.10 × 10-6 6.3082 6 12.60 Yes Exponential 0.390 5 1993 EN50= 6.30 × 10 α=3667 6.3104 5 11.07 Yes 5 N50= 6.30 × 10 Mehta 1996 5 r=3.12 × 10-9 4.4256 4 9.488 Yes Exponential 0.352 8 EN50= 2.22 × 10 α=962 4.4259 3 7.815 Yes 8 N50= 2.22 × 10 Jorgensen 1977 5 r= 9.26 × 10-9 0.4319 4 9.488 Yes Exponential 0.989 b 7 1 EN50= 7.49 × 10 α= 3627 0.4321 3 7.815 Yes 7 N50= 7.49 × 10 Jorgensen 1977 5 r=7.88 × 10-8 0.2924 4 9.488 Yes Exponential 0.990 b 6 2 EN50= 8.79 × 10 α= 291.6 0.2933 3 7.815 Yes 6 N50= 8.79 × 10 Jorgensen 1977 5 r=1.65 × 10-8 2.2233 4 9.488 Yes Exponential 0.695 b 7 3 EN50= 4.19 × 10 α= 0.541 1.4943 3 7.815 Yes 7 N50= 1.88 × 10 Yangco et al. 3 r=4.87 × 10-9 93.975 2 5.992 No Beta-Poisson 0.504 8 1989 EN50= 1.42 × 10 (Disseminated α= 0.201 0.4463 1 3.842 Yes 7 infection) N50= 3.51 × 10 Yangco et al. 3 r= 8.74 × 10-10 18.915 2 5.992 No Beta-Poisson 0.425 8 1989 (Liver EN50= 7.93 × 10 infection) α= 0.048 0.6374 1 3.842 Yes 12 N50= 4.13× 10 a b EN50 is used to denote the median dose for positive endpoint in the Exponential model, analogous to the N50 of the Beta-Poisson model. See Table 5.2 footnote for descriptions of Jorgensen 1977 endpoints corresponding to 1-3. cFail to reject the null hypothesis of acceptability of good model fit (deviance less than the 95% confidence value for the 2 distribution with degrees of freedom equal to the number of doses minus the number of parameters in the model) for p > 0.05.
99
Figure 5.2. Exponential model fit for Tomioka 1993. Pooled M. avium lung lesions serovar 1 strains N-289, N-364, N-445, N-458, N-461 and serovar 9 strains N-254, N- 302 100
Figure 5.3. Exponential model fit for Mehta 1996 for MAC 101- death at 30 days 101
Figure 5.4. Exponential model fit for Jorgensen 1977 (Table 5.3 No. 1)- M. avium serotype 2 (pig origin)- 1. Lymph node lesions: mandibular, parotideus, retropharyng. lat., retropharyng. med., cervicalis sup. dors., cervicalis sup. ventr., subiliacus, popliteus, inguinalis prof., inguinalis superf., spleen. Lymph node Culture: cervicalis sup., cervicalis sup. ventr., subiliacus, ingninalis prof., igninalis superf.liver, kidney 102
Figure 5.5. Exponential model fit for Jorgensen 1977 (Table 5.3 No. 2)- M. avium serotype 2 (pig origin) Lymph node Lesions: Liver, Intestinal mucosa (Peyer Patch) 103
Figure 5.6. Exponential model fit for Jorgensen 1977 (Table 5.3 No. 3) M. avium serotype 2 Culture from retropharyng. lat. lymph nodes 104
Figure 5.7. Yangco 1989 Beta poisson model fit to disseminated infection data 105
Figure 5.8. Beta-Poisson model fit to Yangco 1989 data- liver infection 106
To the author’s knowledge, the median lethal dose (LD50 or microbial dose required to kill 50% of a population of test animals) or ID50 (median infectious dose) for models where the isolate was specifically identified as M. avium hominissuis or
M. chimaera have not been reported. Recently, Winthrop et al. dosed three 12-13 year-old rhesus macaques intrabronchially with 5 mL M. avium subsp. hominissuis with 6.8×108 CFU / mL, 6.8×107 CFU / mL, and 6.8×106 CFU / mL, corresponding to total doses of 3.4×109 CFU, 3.4×108 CFU, and 3.4×107 CFU, respectively (Winthrop et al. 2016). Bacteria were cultured from bronchoalveolar lavage fluid of the animal receiving the highest dose, but not the other two animals, indicating a more human- applicable median infectious dose (ID50) could potentially occur at greater than
3.4×108 CFU.
Although strains of MAC from AIDS patients were used for some of the animal experiments reviewed, no data amenable to modelling was available for animal models using hosts with compromised immune statuses. However, a study of macaques infected with the SIVmac251 strain of simian immunodeficiency virus
(SIV) that were not treated with anti-viral or antibiotic agents unintentionally exposed to 10 to 500 CFU / 100 mL (recovered from 11 of 25 water samples) M. avium in drinking water in the New England Primate Centre resulted in 21 of 67 (31%) of macaques developing M. avium infection (Mansfield and Lackner 1997, Mansfield et al. 1995). Infection rates were lower for other SIV strains (1.9 – 6.7%).
Strains were not consistently reported across studies. This factor is known to be important, however. A study of the virulence of 38 strains of human and environmental M. avium and M. intracellulare isolates in mice demonstrated that human isolates of M. intracellulare were the most virulent followed by environmental isolates of M. intracellulare, human isolates of M. avium, and environmental isolates 107 of M. avium (Tomioka et al. 1993a). This is consistent with a more severe manifestation of lung disease shown in patients with M. intracellulare compared to M. avium, although M. avium infection is more common (Han et al. 2005, Koh et al.
2006, Koh et al. 2012, Tateishi et al. 2009). Pedrosa et al. (1994) conducted a similar study on 41 MAC strains in naturally susceptible mice as well as cultured bone marrow-derived macrophages. They reported strains derived from naturally-infected wild animals were more virulent than AIDS patient-derived isolates or other environmental strains. M. chimaera is therefore thought to be less virulent than M. avium and M. intracellulare, which have more similar virulence (Boyle et al. 2015).
5.11. Epidemiologic support for dose-response models
Where animal models are not available, and/or to provide a “reality check” comparison for modelled median dose values obtained in dose response modeling of animal models, an analysis of attack rates from naturally occurring outbreaks where
(average) dose information is available can be illustrative (Haas 2015). Due to their chronic nature, outbreaks of MAC are not typically reported, with the exception of
MAC hypersensitivity pneumonitis (HP) outbreaks. A dose response model for HP using in vivo whole animal models could not be developed. However, summarizing these outbreaks can also help to contextualize HP-dose relationships.
5.11.6. Hypersensitivity pneumonitis outbreaks
Multiple studies have described point source-associated cases or outbreaks of hypersensitivity pneumonitis after exposure to MAC in hot tubs or spas (Aksamit
2003, Cappelluti et al. 2003, Embil et al. 1997, Fjällbrant et al. 2013, Glazer et al.
2007, Hanak et al. 2006, Kahana et al. 1997, Khoor et al. 2001, Koschel 2006, Lumb et al. 2004, Mangione et al. 2001, Marchetti et al. 2004, Marras et al. 2005, Moraga-
108
McHaley et al. 2013, O'Neil et al. 2006, Pham et al. 2003, Rickman et al. 2002, Sood et al. 2007, Sugita et al. 2000, Travaline and Kelsen 2003). Others hypersensitivity pneumonitis outbreaks have been reported in swimming pool attendants, water damaged-buildings, and occupational settings, but did not identify MAC specifically during patient examinations or environmental investigations (Falkinham 2003). Few
(Fjällbrant et al. 2013, Glazer et al. 2008, Glazer et al. 2007, Lumb et al. 2004,
Moraga-McHaley et al. 2013) have quantified MAC in the water or air that the patient was exposed to, and only three studies (Fjällbrant et al. 2013, Lumb et al. 2004,
Moraga-McHaley et al. 2013) reported concentrations of MAC (note Glazer et al.
2007, 2008 (Glazer et al. 2008, Glazer et al. 2007) reported NTM or slow-growing mycobacteria up to 6 × 105 CFU / mL hot tub water and > 9424 CFU / m3 air above hot tubs). Due to the limited information available on quantitative concentrations of
MAC in water or air paired with disease estimates, a short summary of these three studies is provided below.
Fjällbrant et al. (Fjällbrant et al. 2013) reported a cluster of seven confirmed, probable, and possible hot-tub lung cases with occupational associations to hotel hot tub facilities. Inadequate ventilation, wet storage of filters, and water aerosolization during cleaning were risk factors for the cases. Nonquantitative culture determined that M. avium was present in all hot tubs, but concentrations of M. avium was only determined at one facility, ranging from 140 to 16,000 CFU / mL (Fjällbrant et al.
2013). Two cases were associated with the facility from which M. avium was quantified who had maintained the hot tub at the facility for 8 years. The total number of individuals potentially exposed at that facility was not reported.
Lumb et al. (Lumb et al. 2004) reported three cases of MAC lung disorders in two poorly maintained spa pools with inadequate disinfection. The first two cases
109 were spouses whose isolates matched water isolates from using PFGE. The concentration recovered from their hot tub was 4.3 × 104 CFU / mL MAC. The third case was a 39-year old male who owned a hot tub. His wife and son rarely used the spa pool. The concentration of MAC in their hot tub water was 4.5 × 103 CFU / mL.
An additional case associated with another spa pool was positive for MAC and also
Pseudomonas aeruginosa. The exact number of exposed persons at each hot tub was not reported.
The best study available for providing a quantitative context for HP-dose relationships was prompted by two cases of MAC hypersensitivity pneumonitis in spa workers in New Mexico. Moraga- McHaley et al. (Moraga-McHaley et al. 2013) conducted a study of 56 current (78% of total establishment workforce) spa workers and 1 former employee, noting an exposure-dependent relationship for hypersensitivity pneumonitis-like symptoms. Spa employees with the highest exposure (spa tub cleaner) to aerosolized water reported more symptoms compared to non-exposed workers (odds ratio= 9.6, 95% CI: 1.5-72.7, including index cases).
Intermediately- exposed workers (spa tub workers) also had a higher odds ratio of respiratory symptoms compared to non-exposed workers (odds ratio = 6.5, 95% CI:
1.3-42.3). Isolates from water were identified as M. avium by high-pressure liquid chromatography, polymerase restriction analysis, and 16S rRNA sequencing and water concentrations were all > 800 CFU/ mL (Moraga-McHaley et al. 2013).
Considering only surveyed workers as the exposed populations, attack rates for respiratory symptoms in tub cleaners (n=13), tub workers (n=23), and the non- exposed population (n=21) were 22.8%, 40.4%, and 36.8%, respectively.
Considering only the two hypersensitivity pneumonitis index cases as a numerator and total aerosol exposure workers (n=36), the attack rate was 2/36 or
110
5.6% for a water concentration of > 800 CFU / mL. Specific time-activity patterns and total exposure time of those reporting symptoms were not presented; therefore it is not possible to calculate an incidence rate or person-time rate. This would be a more accurate approach as tub cleaners are likely to be exposed to higher doses for shorter durations than general tub workers who might be exposed to lower doses for longer time durations. Tub cleaners reported the greatest number of mean symptoms per worker (2.31) compared to tub workers (1.26) or non-exposed workers (0.29)
(Moraga-McHaley et al. 2013). A more in-depth dose reconstruction is necessary to make a comparison of exposure dose with an attack rate of 5.6% for >800/mL CFU in hot tub water reported by Moraga-McHaley et al. (2013).
5.11.7. Other MAC infection and disease cases with quantitative dose information
Other studies that conducted environmental investigations and attempted matching of MAC patient isolates and exposure media with or without relation to an outbreak are reviewed in detail elsewhere (Gubler et al. 1992, Halstrom et al. 2015).
Among the studies summarized in these previous reviews for non-HP health endpoints, few provided quantitative estimates of environmental MAC or NTM in conjunction with patient diagnoses.
M. avium (2 CFU / mL) and M. intracellulare (2 CFU / mL) was isolated from a showerhead from a woman with MAC pulmonary disease lacking known risk factors (Falkinham III et al. 2008). In the suspended sediment sample, 240 CFU / mL suspended sediment of M. avium and M. intracellulare were enumerated. Using these data, Pavlik et al. (Pavlik et al. 2009d) estimated that for a 5 minute shower, an individual would be exposed to a total of 3.8 × 104 CFU of mycobacteria. The number of people in the household of the MAC pulmonary disease patient exposed to shower water is not known.
111
In a study of MAC lung disease associated with potting soils (De Groote et al.
2006), aerosol samples were obtained by dropping patient soil samples from a height of 30 cm and measuring Mycobacterium spp. concentrations over a 10-minute period with a 6-stage Anderson sampler operated at a flow rate of 28.3 L per minute, considered to be representative of a human breathing rate. Air concentrations for M. avium ranged from 0.03- 8.59 CFU / 10 L air and M. intracellulare concentrations ranged from 0.03 – 83.96 CFU / 10L air. However, the concentrations of MAC in the initial soil were not quantified. Performing a point estimate calculation with these data, Pavlik et al. (Pavlik et al. 2009d) concluded that this would result in a dose of between 10 and 2400 CFU for the elderly patients infected with mycobacteria from gardening activities. Again, the total number of people exposed to potting soil that resulted in their diagnosis as a MAC patient is not known.
Falkinham 2011 (Falkinham III 2011) found NTM in the household plumbing of 22/37 residences (31 patients) with M. avium (9), M. intracellulare (6), MAC (11),
M. abscessus (4), and M. xenopi (1) isolates. The highest concentrations of NTM were recovered from biofilms (10,371 CFU / cm2), and lesser concentrations from filters
(1,987 CFU / cm2), soils (1,500 CFU / g), and water (157 CFU / mL). The total number of residents exposed to the premise plumbing with those concentrations is not known.
Regarding patients with immune deficiencies, 5/14 (46%) patients under 10 years with an Interleukin-12 (IL-12) deficiency had M. avium infections in a study of otherwise healthy individuals with mycobacterial infections (Altare et al. 1998).
These patients appear to have a lower probability of contracting M. avium infections than HIV/AIDS patients (Pavlik et al. 2009d). According to Pavlik et al. (Pavlik et al.
2009d) one-third of individuals with IL-12 deficiency exposed to a cumulative dose of
112
3.5 × 104- 2 × 106 CFU of M. avium through water would develop M. avium bacteremia.
8/19 (42%) of patients with pulmonary alveolar proteinosis (and all smokers) were infected with M. avium (Witty et al. 1994). Pavlik et al. (Pavlik et al. 2009d) estimates that for a 10-year exposure to aerosols containing MAC, with an inhalation rate of 6 m3 / day and 3000-fold average enrichment of MAC aerosols compared to bulk water (Parker et al. 1983), 42% of patients with pulmonary alveolar proteinosis exposed to a dose of 2.2 × 105- 1.1 × 107 CFU M. avium would develop M. avium pulmonary disease. However, the authors noted that in addition to environmental sources, cigarettes can contain M. avium (Eaton et al. 1995a).
The summarized findings indicate that while current human epidemiologic data does not provide sufficient estimates of the total number of people exposed for outbreaks and environment-patient matching studies reported in the literature, it can be used for comparison with the derived dose response models. Estimates of the dose inhaled by a pulmonary disease patient exposed during showering (3.8 × 104 CFU)
(Pavlik et al. 2009d) are comparable with the N50 calculated for lung lesions (6.30 ×
105 CFU) in the Tomioka et al. lung lesions model within an order of magnitude.
Similar estimates for the dose inhaled MAC from potting soil handling by elderly patients with pulmonary mycobacterial infections (10-2,400 CFU) are 2-4 orders of magnitude lower than the Tomioka model N50. This could be due to the fact that the patients in the soil study (de Groote et al. 2006) were chosen for inclusion based on infection rather than a more serious presentation of lung infection. A significant proportion of the elderly individuals included in the soil study were gaunt or had mutations in either the CFTR or α-1-antitrypsin genes. Underestimation of soil inhalation in the Pavlik et al. point estimate calculation, or selection for an elderly
113 disease population in the study design could also play a role. Pavlik et al. also estimated that 42% of patients with pulmonary alveolar proteinosis who were smokers were exposed to a dose of 2.2 × 105 – 1.1 × 107 CFU. In the Tomioka model fit, a
42% infection rate was associated with a median and 95th percentile dose of 4.9 ×105
CFU and 3.4 × 105 CFU, respectively. Pavlik’s estimate is higher than the modeled
Tomioka et al. doses, but was estimated over a 10 year time period rather than a single exposure.
For SIV-infected macaques (see section 4.1) (Mansfield and Lackner 1997,
Mansfield et al. 1995), a 31% infection rate was associated with a dose of 1 × 104 and
5 × 105 CFU (Pavlik et al. 2009d). A 31% probability of disseminated infection is associated with a median and 95th percentile dose of 6.1 × 106, 5.8 × 105 CFU for the
Yangco et al. model for disseminated infection, approximately one order of magnitude higher than the Pavlik estimate. For disseminated infection in immune- compromised patients, based on assumptions derived by Pavlik (2009d) from the
Mansfield studies, ~33% of patients with IL-2 deficiency would develop M. avium bacteremia when exposed to doses of 3.5 × 104- 2 × 106 CFU. A 33% probability of disseminated infection is associated with a median, 95th percentile dose of 6.8 × 106
CFU, 8.5 × 105 CFU for the Yangco et al. model for disseminated infection. Both comparisons support that the Yangco dataset could be a reasonable model for disseminated infection in immune-compromised patients.
5.12. Incorporation of dose response models into the QMRA framework
5.12.6. Selecting an appropriate dose response model for a particular population and health endpoint
114
The QMRA framework proposed indicates that for different exposure groups and host categories, we can identify the predominating MAC strains and use dose response models that apply, along with conditional probabilities for disease progression. For pulmonary infection, the models derived from Tomioka et al. and
Mehta et al. can be applied for sub-clinical/moderate and clinical severity lung infection, respectively, as has been applied in Legionella QMRA models using animal dose response models with infection or death endpoints by Armstrong et al.
(Armstrong and Haas 2007a, Armstrong and Haas 2007b). The disseminated infection model from Yangco et al. can be used for the development of disseminated infection in HIV/AIDS patients, while Jorgensen et al. models can be used to model cervical lymphadenitis in children. Regarding hypersensitivity pneumonitis, a dose response model could not be developed, but the best information regarding occupational exposure in healthy adults indicates an attack rate of 5.6% at MAC water concentrations of > 800 CFU / mL. The physics of QMRA models for various scenarios like showering and use of the functional forms (Exponential, Beta-Poisson) of the dose response models discussed here will be the same as the mathematical models previously reported for other pathogens like Legionella for common water uses (Schoen et al. 2011, Hamilton and Haas 2016)(Hines et al. 2014b).
5.12.7. Extrapolating from experimental exposure routes
In many cases, in vivo animal experiments do not use the same route for dosing animals that might be representative of a human exposure scenario (for example, inhalation of a shower mist is probably a less potent exposure route than being directly injected intravenously as an animal might have been during in an experiment). Therefore, there are limitations for QMRA modeling with utilizing dose- response models derived from an intravenous route when the desired modelled
115 exposure route is inhalation or oral. To address this gap explicitly in a QMRA model, theoretically, a conversion factor or distribution of conversion factors could be introduced to the model as a parameter to extrapolate from the experimental exposure route to the desired modelled human exposure route. Limited information is available for comparing routes of exposure. The LD50 for M. avium 724 is reported as > 5 ×
108 CFU by the intravenous (IV) route, and > 1 × 106 by the inhalation route at 90 days (Collins 1985). This LD50 is aligned with our model results from the Mehta et al. (Mehta 1996) intravenous dataset (LD50 2.22 × 108). This factor of 500 between the intravenous and inhalation route could potentially be used to convert between exposure routes within the context of a QMRA.
The LD50’s for M. avium 702, M. intracellulare 1403, and M. intracellulare D673 are reported as 4 × 108 (intravenous), > 5 × 108, and 5 × 107, respectively (Collins et al. 1978). One study (Olitzki et al. 1968) reports the LD33 for intracerebral infection with M. avium serotype I (strain 2827-536, chicken origin as per Schaefer seroagglutination criteria that are now known to be ambiguous for identifying specific species or sub-species of MAC (Schaefer 1965, Schaefer et al. 1970, Schaefer 1967,
1968, Wayne et al. 1993)) as between 1 × 108 and 2 × 108 CFU and an LD50 for serotype II (strain 1035-496, origin not specified) greater than 1.2 × 108 CFU.
Olitzki et al. further state that “…in order to obtain a generalized infection of all big organs (spleen, liver and lung) it suffices to inject 670 microorganisms of strain
M. avium, type II [strain 1035-496], intracerebrally.” The minimal effective doses to achieve specific tuberculous processes with the appearance of epithelioid tubercles in
6 7 5 the majority of organs were defined as 2 × 10 CFU, 2 × 10 CFU, 6.7 ×10 CFU, and
2 × 105 CFU for intraperitoneal, subcutaneous, intracerebral, and intravenous routes, respectively. The phrase “minimal infectious dose” is historically applied to represent
116 the dose at which no positives are observed, or can also be a misnomer applied to connote the ID50. If a conversion factor of 3.25 is applied between the intracerebral and intravenous route, this would indicate an LD50 of > 4.2×108 CFU for serotype II, and ID50 for specific tuberculous processes and disseminated infection of 2×105 CFU and 2.4×103 CFU, respectively. These values are comparable with those modelled in the current study.
Inhalation studies (Olitzki et al. 1968) were also performed by the same authors using a 1- hour aerosol exposure at 50% relative humidity and 23.3ºC, but an infective dose was not reported. For dosing inoculums ranging from 1.5 × 108 - 1 ×
109 CFU, 3 × 103 - 4.2 × 104 CFU were recovered from mice lungs at 20 hours post inoculation, indicating colonization and/or infection at these doses. However, the number of animals challenged in each dose group was not specified which prevented further dose-response modeling or comparisons. It is noted that in most cases, it was difficult to equate historically reported strain numbers with current MAC taxonomic descriptions, limiting the utility of historic datasets which do not describe the specific origin of the strains.
5.12.8. Conditional probabilities for disease outcomes
Within the context of QMRA, conditional probabilities are frequently applied to translate microbial infection to disease pathogenesis where for example, only an infection dose response model is available when an endpoint of disease is desired.
Movement from one state (for example infected) to a disease state (such as pulmonary disease) can be described by multiplying the probability of transition to one state by the probability of being within the other state, i.e., Pdisease = Pinfection*Pdisease|infection.
The proportion of patients with MAC in their sputum whom will eventually acquire lung disease is unknown and likely to be rare, making it difficult to develop
117 conditional probabilities of disease given infection for application in a QMRA
(Inderlied et al. 1993). Predisposing lung conditions, body habitus, genetic predisposition, immune factors, and ciliary abnormalities play a role in disease development (Field et al. 2004, McShane and Glassroth 2015). Furthermore, the impact of early treatment on this disease progression is also not well characterized
(Field et al. 2004). Fibrocavitary MAC disease is typically treated immediately due to the potential for rapid lung damage (Griffith et al. 2007). Nodular bronchiectasis displays more variable disease progression, and patients without apparent lung disease may be colonized with MAC (Daley and Griffith 2010). However, the introduction of high-resolution CT scans has demonstrated the slow, sometimes decades-long progression of the fibronodular bronchiectasis form of MAC lung disease could warrant treatment to prevent the development of lung disease even in the absence of symptoms, suggesting that QMRA might still consider these colonized individuals at risk (Field et al. 2004).
Studies of antibody and skin test reactivity provide information on exposure.
Theoretically, these methods or another biomarker of exposure and could potentially be used to calculate transition probabilities of MAC infection to MAC disease by calculating (number of MAC severe infection or disease cases in a given population over a period of time) / (the number of people with positive antigen tests i.e. number exposed over the same time period). However, sensitization to mycobacterial antigens can also be provoked by the Calmette-Guérin (BCG) vaccine (Farhat et al. 2006), and because MAC infections are not nationally notifiable, time periods of recorded disease and sensitization typically do not overlap. Studies of mycobacterial exposure have shown high levels of mycobacterial skin reactivity and lack of NTM-associated disease outcomes in the studied populations and/or have targeted healthy or disease
118 populations for testing rather than a random sample where both antigen and disease status were determined (Abrahams and Harland 1968, Bardana et al. 1973, Bermudez et al. 1989, Kwamanga et al. 1995, Lee et al. 1991, Shachor et al. 1997, Thayer et al.
1990, von Reyn C et al. 2001, von Reyn et al. 1993a). However, according to a survey of PPD-B (M. intracellulare) skin reactivity of 18- to 25- year old men in the southeastern coastal United States, greater than 60% showed a positive skin reaction, indicating they are, or at some stage were colonized or infected but did not show signs of disease (Edwards et al. 1969). Another more recent study (Khan et al. 2007) demonstrated that from 1971-1972, one in nine (~11%) people in the United States noninstitutionalized civilian population were sensitized to M. intracellulare while one in six (~17%) were sensitized from 1999-2000 cohorts. These findings support that potentially only a small portion of MAC exposures may progress to severe infection or disease outcomes in a healthy population (Primm et al. 2004), but the portion for
HIV patients is likely to be larger (Arasteh et al. 2000) and exposures to MAC have been shown to increase with age (Fairchok et al. 1995). Adjemian et al. (2012b) estimated the prevalence of NTM lung disease in US medicare beneficiaries (>65 years of age) to be 20 cases per 100,000 persons in 1997 and 47 cases per 100,000 persons in 2007 (8.2% increase per year). If we assume that 80% of NTM cases are due to MAC (Prevots et al. 2010), this would equate to 37.6 cases per 100,000 persons in 2007. The 2007 population over age 65 was approximately 37.3 million people in 2007 (US Census 2007), resulting in an estimate of 14,025 cases in 2007.
Using the same population size of 37.3 million, a 1 in 6 exposed population (Khan et al.) would equate to 6.22 million people exposed. The probability of MAC pulmonary disease in a > 65 years of age population given MAC exposure would then be approximately 14,025 cases / 6.22 × 106 people exposed or 0.226%. There are
119 numerous limitations of this approach including comparing populations of differing age ranges; however, this is demonstrative of the type of calculations that can be performed to estimate the probability of MAC illness given a MAC infection
(Pdisease|infection).
An alternative approach can use case-control studies of MAC disease patients and matched controls, or longitudinal studies of MAC seropositive patients. These data can be combined with reported disease rates in the general population to estimate transition probabilities indirectly using Bayes rule as has previously been performed in another pathogen of complex pathogenesis, Helicobacter pylori (Ryan et al. 2014).
As longitudinal studies of disease incidence in MAC seropositive populations have not been performed, for each MAC disease endpoint, the case-control approach would involve determining the proportion of patients who were seropositive for MAC in a disease outcome group versus a matched control group. To the authors’ knowledge, insufficient information is available to conduct such an analysis at this time.
5.13. Limitations and research gaps
5.13.6. Variation in virulence between MAC species
Although knowledge of dominant strains of MAC can aid in choosing the appropriate dose response model and conducting a risk assessment, MAC epidemiology may continue to change. In in recent decades, it has been suggested that the slightly higher virulence species M. intracellulare (Birkness et al. 1999, Boyle et al. 2015, Weiss and Glassroth 2012) is representing a more important portion of the total MAC disease burden compared to M. avium (Carter et al. 2014, Falkinham III
2016, Thomson 2010, Wolinsky 1979). This is reflected in transitioning
120 epidemiological patters from cavitary disease in older men with pre-existing conditions and/or occupational exposures to bronchiectasis in healthy older women and could explain an increased proliferation of MAC diagnoses overall (Falkinham III
2016). Patients with M. intracellulare infection are also more likely than those with
M. avium to suffer a reinfection or relapse (Boyle et al. 2015). However, the future directions of MAC epidemiology and their effect on the applicability of the proposed
QMRA framework are unknown. As M. chimaera is the most recently differentiated species (Tortoli et al. 2004), more work is needed to decipher differences in exposure and clinical outcomes related to this MAC species. Due to changing MAC taxonomy, the interpretation of MAC species over time is not straightforward. Increased attention to speciation and clear reporting of MAC species in both the clinical and environmental research community, especially with regard to in vivo animal models, would prove quite useful.
5.13.7. Impact of environmental conditions on virulence
Disinfection processes may be one reason for this epidemiologic shift, for example in Australia, increased drought and/or water restrictions may have led to increased chlorine degradation time and lower disinfectant residual at the point of use
(Thomson 2010). Temperature increases can also degrade chlorine faster, further decreasing disinfectant residuals. Alternatively, better disinfection favors environmental NTM, specifically those that are able to better sequester themselves within biofilms (Falkinham III 2016, Norton and LeChevallier 2000a). Both M. avium and M. intracellulare are found more frequently on pipe surfaces compared to bulk water (Falkinham III et al. 2001, Torvinen et al. 2004b), however, M. intracellulare forms biofilms more readily than M. avium, and M. avium is more commonly found in the bulk water than M. intracellulare (Falkinham III et al. 2001, Kirschner Jr et al.
121
1992). The high cell surface hydrophobicity of mycobacteria likely results in their preferential attachment to surfaces and air-water interfaces (Van Oss et al. 1975); this hydrophobicity varies between MAC species (Stormer and Falkinham 1989). M. intracellulare is slightly more hydrophobic and accumulates more readily on the surface of glass beads than M. avium (Steed and Falkinham III 2006). One study noted that switching from chlorine to chloramines favored an increase in
Mycobacterium spp. in biofilms (Pryor et al. 2004b). This may contribute to MAC proliferation as utilities experience growing pressure to decrease disinfection byproducts. Clinical isolates of MAC are more likely than environmental isolates to display resistance markers or contain plasmids; furthermore, aggregate forms of M. avium subsp. hominissuis are more virulent than the planktonic form (Babrak et al.
2015, Lopes Leivas Leite 2015), emphasizing the importance of biofilms in water systems as modulators of virulence. Pavlik et al. (2009b) describe environmental factors affecting the virulence of M. avium subsp. avium as the external environment of the host organism (presence of inhibitory agents and antibiotics, ultraviolet radiation), and the internal environment of a host organism (protozoa or vertebrate tissues). Work is needed to assess how environmental dynamics influence colonization of MAC within biofilms, and therefore health risks. A control strategy for MAC must address biofilm dynamics as “the only [NTM] cells in water are those released from the biofilm” (Pavlik et al. 2009c).
5.13.8. Impact of microbial ecology on virulence
In particular, the relationship between MAC and free-living amoeba (FLA) has been shown to play a role in virulence. The ability for multiple MAC species to survive and replicate in both the trophozoite and cyst state of the FLA species Acanthamoeba polyphaga, Acanthamoeba castellani, Dictyostelium discoideum, and Tetrahymena
122 pyriformis has been demonstrated (Adékambi et al. 2006, Thomas and McDonnell
2007). M. avium has been shown to replicate more efficiently in FLA than other NTM such as M. marinum, M.fortuitum, and M. smegmatis, further indicating that MAC is a useful model for FLA-NTM interactions in premise plumbing (Cirillo et al. 1997).
Although the dynamics of FLA-bacteria interactions have been explored somewhat for L. pneumophila (Buse and Ashbolt 2011, Buse and Ashbolt 2012, Kuiper et al.
2004b), the mechanisms and biological significance of the interactions between MAC and FLA are not well characterized. MAC are thought to use similar mechanisms to
Legionella spp. for inhibiting fusion of the FLA’s phagosome and its bacteria- degrading lysosome, however, their location in the cell is fundamentally different from Legionella containing vacuoles (LCV), and instead has been identified as the cell exocyst under laboratory conditions (Salah and Drancourt 2010). These rates and mechanisms of phagocytosis, survival, and growth of MAC by FLA are further likely to differ from those of Legionella spp. due to differences in cell surface hydrophobicity, and considerable variation in FLA interactions within the NTM group itself(Drancourt 2014, Mura et al. 2006, Strahl et al. 2001). Furthermore, there have been no assessments of mycobacterial growth in protists in natural habitats to date
(Pavlik et al. 2009a).
5.13.9. Animal models
Seven new exponential dose response models have been proposed here for human-relevant species of MAC with endpoints of lung lesions, disseminated infection, liver infection, death, lymph node lesions, and lymph node infection in experimental in vivo animal models. Current physical and biochemical tests used in clinical settings do not typically differentiate between M. avium and M. intracellulare due to its limited subsequent prognostic and therapeutic advantages (Field et al. 2004,
123
Griffith et al. 2007, Kasperbauer and Daley 2008). However, differentiating between environmental species and subspecies of the MAC can aid in the assessment of health risks, development of improved laboratory detection methodologies, more tailored diagnostic and treatment approaches, and control of MAC sources (Griffith et al.
2015, Tsang et al. 1992, Turenne et al. 2007). Furthermore, this information is useful for conducting a QMRA.
It is noted that for the lung lesions model (Figure 5.2), the exponential model provided an acceptable statistical fit; however the 99% confidence interval did not encapsulate two data-points. Further exploration of other mechanistic models is necessary to improve model fit for lung lesions by adding additional parameters. One such model is a multi-hit dose response model, which will result in a “steeper” curve
(Haas et al. 1999).
All seven models derived in this work are for M. avium or MAC from in vivo studies. However, in vitro studies and cell culture methodologies have been used to assess infectivity for other pathogens such as Cryptosporidium (Slifko et al. 1999), and may also present alternatives to animal models that could be explored for MAC.
MAC infection models have generally been performed in macrophages (mouse peritoneal macrophages, mouse alveolar macrophages, mouse bone marrow-derived macrophages, human blood monocyte-derived macrophages, and human alvelolar macrophages from HIV patients) to assess the antimicrobial activity of chemotherapeutic agents or virulence (Pedrosa et al. 1994, Perronne 1998). Animal models for M. avium subsp. hominissuis currently under development using the nematode Caenorhabditis elegans may prove useful for future assessments of host- pathogen interactions (Everman et al. 2015).
124
Additionally, although several of the proposed dose response models utilized
MAC strains from AIDS patients during testing, no data amenable to modeling was available from studies with immunocompromised animal hosts. It has been demonstrated the dominant strain of MAC bacteria varies for various health outcomes. However, here we have not addressed all MAC species important to MAC disease, nor all potential health outcomes or underlying susceptibilities affecting
MAC exposure or infection. For example, we have not addressed allergic reactions provoked by environmental mycobacteria such as non-specific reactions to tuberculins, cutaneous basophilic hypersensitivity, or allergic reactions of the digestive tract (shown to occur at an effective dose of 106 M. avium subsp. hominissuis in guinea pigs and rabbits) (Pavlik et al. 2009d). Quantification of the relative importance of host disease status, exposure route, and MAC species on clinical outcomes is a key area for future work.
5.14. Conclusions
The proposed risk framework elucidates the complexities resulting from exposure route, host susceptibility, and strain difference characteristics. This diagram can be used with the proposed dose response models to conduct quantitative health risk analyses for environmental and food-borne MAC. In addition, we anticipate that the concepts presented could form the basis for a Key Events Dose-Response Framework
(KEDRF) based on “mode-of-action” (MOA) concepts or fundamental biological events and processes that occur between initial exposure and the health endpoint of concern (Julien et al. 2009). This approach seeks to characterize the kinetics and dynamics of each key event. The interplay of dose, “control mechanisms” (specific mechanisms that may influence the ultimate outcome), and host characteristics (life
125 stage, disease state, genetic makeup, exposure patterns). Importantly, the proposed dose response models for MAC provide a quantitative description of the interaction between MAC pathogens and hosts, a dynamic that is highly complex. This information is necessary for analyses that will inform engineering and public health risk-mitigation decisions, and help to prioritize research needs in this area.
126
6. Quantitative microbial risk assessment (QMRA) of Legionella and MAC in roof-harvested rainwater4
6.1. Abstract
A quantitative microbial risk assessment (QMRA) of opportunistic pathogens Legionella pneumophila (LP) and Mycobacterium avium complex (MAC) was undertaken for various uses of roof-harvested rainwater (RHRW) reported in Brisbane, Australia to identify appropriate usages and guide risk management practices. Risks from inhalation of aerosols due to showering, swimming in topped up pools, use of a garden hose, car washing, and toilet flushing with RHRW were considered for LP while both ingestion
(drinking, produce consumption, and accidental ingestion from various activities) and inhalation risks were considered for MAC. The drinking water route of exposure presented the greatest risks due to cervical lymphadenitis and disseminated infection health endpoints for susceptible populations including children and the immune-deficient, respectively. It is therefore not recommended that these populations consume untreated rainwater. LP risks were up to 5 orders of magnitude higher than MAC risks for the inhalation route of exposure for all scenarios. Both inhalation and ingestion QMRA simulations support that while showering and use of high-pressure sprays for car washing may present risks in excess of a 10-4 annual infection risk reference point, toilet flushing would constitute an appropriate use of RHRW.
4 This chapter is under review at a journal
127
6.2. Keywords
Roof-harvested rainwater; opportunistic pathogens; quantitative microbial risk assessment (QMRA); Legionella pneumophila; Mycobacterium avium complex;
6.3. Introduction
Globally, harvested rainwater is used to supplement both potable and non-potable water supplies. In Australia, roof-harvested rainwater (RHRW) constitutes an important source of water for many households; in 2010, 32% of Australian households had a rainwater tank and rainwater tanks were the main source of drinking water for 13.6% of
Queensland households (ABS 2010). Queensland rainwater tank owners have reported numerous potable and non-potable uses for their RHRW, including drinking, cooking, clothes washing, showering, pool top-up, gardening, car washing, ornamental water features, toilet flushing, filling fish tanks, and pet washing (Hamilton et al. 2016). This indicates the potential for exposure to rainwater through numerous scenarios.
Previous studies have identified high concentrations of both enteric pathogens such as Salmonella, E. coli, Campylobacter spp., Cryptosporidium spp., Giardia spp. (Ahmed et al. 2011b, Ahmed et al. 2010, Crabtree 1996) and opportunistic pathogens such as LP,
MAC, Aeromonas hydrophila, Staphylococcus aureus, Pseudomonas aeruginosa,
Acanthamoeba spp. (Ahmed et al. 2014b, Hamilton et al. 2016) in Queensland rainwater tanks. Opportunistic pathogens cause illness primarily in individuals with underlying health conditions, children, and/or the elderly. However, they are a growing cause of drinking water-associated disease outbreaks worldwide (Falkinham et al. 2015).
128
There is epidemiologic evidence of disease cases associated with enteric pathogens
(Brodribb et al. 1995, Franklin et al. 2009, Koplan 1978, Merritt et al. 1999, Murrell and
Stewart 1983, Simmons and Smith 1997) as well as opportunistic pathogens LP (Schlech
III et al. 1985, Simmons et al. 2008) and MAC (Lumb et al. 2004) in RHRW. While drinking water guidelines used to determine the safety of Australian rainwater specify the non-detection of E. coli in 100 mL of water (NHMRC-NRMMC 2004, WHO 2004), there is no consensus about which end-uses of rainwater are appropriate with regards to opportunistic pathogen associated health risks. No study has assessed the full suite of potential rainwater uses for an opportunistic pathogen to make such a determination. This is especially important as treatment options are limited for rainwater tank owners and are typically limited to gutter protection, first-flush devices, or point-of use filters. These options have limited efficacy for removing pathogens. (Dobrowsky et al. 2015a,
Dobrowsky et al. 2015b, Egodawatta et al. 2009, Jordan 2008, Kus et al. 2010, Mendez et al. 2011, Reyneke et al. 2016).
Quantitative microbial risk assessment (QMRA) can be used for the purposes of estimating the human health risk associated with exposure to pathogens in environmental media using a process of hazard identification, exposure assessment, dose response, and risk characterization (Haas et al. 1999). Due to their direct linkage to RHRW-associated disease cases (Lumb et al. 2004, Simmons et al. 2008), global epidemiologic importance
(Falkinham 3rd et al. 2015), and known occurrence in RHRW (Hamilton et al. 2016), LP and MAC were chosen as index opportunistc pathogens for a QMRA estimate. While there are over 50 species of Legionella and several are human pathogens, LP is the most common species that causes the severe pneumonia-like illness Legionnaires’ Disease, as
129 well as the less severe form of illness, Pontiac fever (Diederen 2008, Muder and Victor
2002). The rate of Legionellosis in Australia was 13 per million people in 2012, and eighty cases were reported in Queensland in 2015 (Australian Government Department of
Health 2016b, Phin et al. 2014). MAC is a subset of non-tuberculous mycobacteria
(NTM) that can cause soft tissue infections and cervical lymphadenitis in immune- competent patients, disseminated infections in immune-compromised patients, and pulmonary disease in both healthy and immune-compromised groups (Falkinham III
1996). MAC were the most common pathogen in NTM isolates in Queensland in 2005, and most frequently identified isolate in NTM cases in the Northern Territory, Australia from 1989-1997 (O'Brien et al. 2000, Thomson et al. 2013c), however NTM-associated illnesses are not reportable in Australia. MAC is comprised of 9 species (M. avium, M. intracellulare, M. arosiense, M. chimaera, M. colombiense, M. marseillense, M. timonense, M. bouchedurhonense, and M. ituriense) (Falkinham III 2013a). The most human-relevant species and therefore the focus of this QMRA are M. avium (comprised of four subspecies: paratuberculosis, avium, hominissuis, and silvaticum), M. intracellulare, and M. chimaera (Hamilton et al. 2017).
Previous QMRA studies of RHRW have focused on enteric pathogens or LP, typically focusing on one or two exposure scenarios (Ahmed et al. 2010, de Man et al.
2014a, Fewtrell and Kay 2007b, Lim et al. 2015, Lim and Jiang 2013, Schoen and
Garland 2015, Schoen et al. 2014). This has been partially due to the lack of dose response models for opportunistic pathogens such as MAC. A single previous assessment of exposure to MAC is available for treated tap water from a centralized distribution system (Rice et al. 2005). However, this study did not quantify health risks. A dose
130 response model has since been developed for one MAC subspecies, M. avium subsp. paratuberculosis (MAP) (Breuninger and Weir 2015), however the relationship between human exposure to this subspecies and the development of health effects (it is postulated that Crohn’s disease may be the health outcome for this pathogen (Pierce 2009, Waddell et al. 2015)) is contentious and is therefore excluded from the current analysis. A family of MAC dose response models was recently developed for human-relevant species of
MAC in environmental media, allowing for development of a population and exposure- route specific QMRA for MAC risks (Hamilton et al. 2017). For LP, generally only inhalation or aspiration routes are considered relevant, with inhalation being the most common exposure route (Ellis 1993).
The goals of the current study are therefore to assess the health risks from index pathogens LP and MAC by conducting a QMRA of multiple potential exposure scenarios with (1) inhalation exposure to LP and (2) inhalation and ingestion exposures to MAC.
6.4. Materials and Methods
6.4.1. Exposure models
In our previous study, Queensland rainwater tank owners reported using RHRW for drinking, cooking, clothes washing, showering, pool top-up, gardening, car washing, ornamental water features, toilet flushing, filling fish tanks, and pet washing (Hamilton et al. 2016) (Appendix Table 9.20). In that study, LP and MAC were measured in 134 rainwater tanks from Brisbane, Australia. The use of RHRW for ornamental water features, filling fish tanks, and pet washing were reported by less than 2% of surveyed
131 residents, and were therefore excluded from the current QMRA analysis. The remaining potential exposure pathways for LP and MAC are shown in Figure 6.1. As reported in a previous study (Hamilton et al. 2017), exposure to MAC species can occur through: (1) the ingestion of food, water (M. avium, M. chimaera), or soil (M. intracellulare); (2) inhalation of water aerosols (M. avium, M. chimaera) or soil dusts (M. intracellulare); (3) aspiration; or (4) iatrogenic exposure. Scenarios (1) and (2) were considered here as it is unlikely to use harvested rainwater in a medical setting, and limited information is available regarding aspiration rates and the amount of pathogenic material transferred to the lungs during each aspiration event. Exposure models for each scenario are summarized below and in Figure 6.1.
132
Figure 6.1 Exposure routes for L. pneumophila and MAC in roof-harvested rainwater
133
6.4.1.1. Ingestion exposure models
As Legionellosis is contracted through inhalation or aspiration, ingestion exposure models apply only to MAC. The general process for ingestion of RHRW for a variety of applications for MAC can be described by equation (6.1). Additional factors and modifications for each scenario are described below and summarized in Table 6.1.
1 퐷 = 퐶 푉 Equation 6.1 푖,푗 푅 푅퐻푅푊,푀퐴퐶 푖푛푔
Where Di,j = The daily dose of of pathogen i (where i = LP or MAC) for exposure scenario j (where j = drinking, eating produce, etc.), CRHRW = Concentration of MAC in
RHRW, R = recovery efficiency (%), and Ving = the volume of RHRW ingested per exposure event.
6.4.1.1.1. Drinking
A mean of 0.869 L per day and 95th percentile 2.717 L per day (USEPA 2011a) was used to construct a lognormal distribution for daily drinking water intake. Two scenarios, without filtration and with filtration (equation 6.2) were considered as some participants in our previous rainwater study (Hamilton et al. 2016) had under-sink point-of-use (POU) filters. Chlorine disinfection was not considered as only one participant of 134 households reported having ever used chlorine to disinfect their rainwater (Hamilton et al. 2016). The most common treatment method for those who reported treating their rainwater was filtration (29/134 participants) (Hamilton et al. 2016). For a POU filter used with rainwater systems, a previous study reported a 39% (0.4 log) reduction in E.
134 coli bacteria (Jordan 2008). A sand filter used with rainwater tanks was associated with a
99% (2 log) removal of bacteria (Ahammed and Meera 2010). A uniform distribution for filtration of 0.4-2 log removals was used.
1 퐷 = 퐶 푉 10−퐿 Equation 6.2 푖,푗 푅 푅퐻푅푊 푖푛푔
Where –L is the number of log removals for filtration.
6.4.1.1.2. Consumption of raw produce
The consumption of uncooked lettuce was chosen as the index scenario for eating produce irrigated with RHRW. The selection of lettuce was due to its high water retention compared to other crops because of its large and uneven surface area, its high probability of being eaten raw, high lettuce consumption compared to other produce crops by Australian populations, and short shelf life that bounds the possible time between harvest and consumption, especially in a subtropical climate (Ahmed et al.
2016). Contamination with MAC in lettuce can occur through processes of: (1) the irrigation water that adheres to the outside of the plant and (2) MAC is internalized into plants (Kaevska et al. 2014). Internalization can occur through uptake through the roots, or through stomata or wounds present on the leaf surface (Hirneisen et al. 2012). These microorganisms would not be washed off by the consumer as they are located inside of the plant. To account for both of these processes (surface-attached and internalized M. avium), the total dose of MAC on the surface of lettuce and the MAC internalized in the
135 plant are summed to arrive at the total dose (equation 6.3 - equation 6.5) (Sales-Ortells et al. 2014).
1 퐷 = 퐶 푉 10−푘푓,푠푡푓10−푊퐼 Equation 6.3 푀퐴퐶,푠푢푟푓푎푐푒 푅 푅퐻푅푊 푅
1 퐷 = 퐶 퐹 푉 퐼 Equation 6.4 푀퐴퐶,푖푛푡푒푟푛푎푙푖푧푒푑 푅 푅퐻푅푊 푖푛푡 푅
퐷푀퐴퐶,푡표푡푎푙 = 퐷푀퐴퐶,푠푢푟푓푎푐푒 +퐷푀퐴퐶,푖푛푡푒푟푛푎푙푖푧푒푑 Equation 6.5
Where VR = volume retained per gram of lettuce during irrigation (L per gram), I
= gram consumed per person per day (grams per person per day), Kf,s = in-field decay constant for MAC on the surface of lettuce (per day), W = log reductions due to washing of lettuce prior to consumption, Kf,int = in-field decay constant for internalized MAC (per day), Fint = the internalized fraction of MAC in the irrigation water that is found in the leaves (in CFU per gram of lettuce), tf = withholding time between harvest and consumption (days). Lettuce is assumed to be immediately consumed by the homeowner after harvest, and, therefore decay of MAC due to transport and/or storage of the lettuce is not considered.
Internalization has typically been reported for enteric bacteria or viruses in a variety of produce (Dicaprio et al. 2012, Solomon et al. 2002, Wei et al. 2011), however, only one report of internalization of M. avium in plants is available. Internalization rates were calculated from Kaevska et al. (Figure 6.2 and 6.3) data from field studies with tomato plant leaves by computing (concentration recovered from leaves after surface washing) / (concentration of seeding inoculum) for lettuce. Internalized fractions ranged
136 from 1.13 × 10-5 to 9.49 × 10-4 for M. avium subsp. paratuberculosis and 1.00 × 10-8 to
1.37 × 10-5 for M. avium subsp. avium. A uniform distribution using the more conservative M. avium subsp. paratuberculosis values was used. During field experiments conducted by Kaevska et al. over 8 weeks, concentrations of M. avium in leaves did not consistently increase or decrease between 3 (M. avium subsp. avium) or 4 (M. avium subsp. paratuberculosis) sets of experiments. For this reason, decay was not included in the model for internalized M. avium.
For the volume of water retained by lettuce during irrigation, Shuval et al. estimate a mean and standard deviation of 0.108 ± 0.019 mL per g for green oak lettuce planted in beds with three staggered rows spaced 30cm apart with fixed-set overhead sprinklers at
10 min intervals. It was assumed that microorganisms are homogeneously distributed in
RHRW and 100% of the organisms in the retained irrigation water become initially attached to the lettuce surface.
Decay experiments for M. avium on the surface of plants were not available, however
Cook et al. (2013)) performed decay experiments for M. avium subsp. paratuberculosis exposed to silage exudates derived from grass and alfalfa. A decay k value of -0.0484 d-1 was reported. Removal of E. coli during washing with water of 0.3 ± 0.1 log (Holvoet et al. 2014) was assumed to be representative of removal of M. avium.
6.4.1.1.3. Showering
A uniform distribution of water accidentally ingested per daily shower event was used of
58 µL - 1.9 mL (Ahmed et al. 2010).
6.4.1.1.4. Gardening
137
Estimates for accidental ingestion during use of a garden hose has been reported to range from 0.002 -1.9 µL (Ahmed et al. 2010) up to 1 mL (NRMMC-EPHC-AHMC 2006) or
1.1 mL (Schoen et al. 2014). Irrigation of lettuce with recycled water was estimated to occur 90 times per year (NRMMC-EPHC-AHMC 2006).
6.4.1.1.5. Car washing
The total volume of water consumed during 10 minutes of car washing using a high- pressure spray device was recently estimated using cyanuric acid as a tracer of water ingestion. Among 26 participants, the accidental ingestion volume per car washing event was estimated to range from 0.06 to 3.79 mL (Sinclair et al. 2016). A monthly car washing frequency was assumed (Villarreal and Dixon 2005).
6.4.1.1.6. Pool top-up
Australian pool-owners reported using RHRW as well as drinking water to fill their swimming pools and all pools observed in the previous study were outdoor family pools
(Hamilton et al. 2016). The distribution of the proportion of drinking water: RHRW is not known as pool sizes, tank sizes, and pool maintenance practices vary considerably. As a result, scenarios of 10%, 50%, and 90% dilution of RHRW with opportunistic pathogen- free water in pools were modelled (equation 6.6). Additional pathogen decay through mixing of RHRW with tap water with a chlorine residual was considered to be minimal and potential pathogen removal due to pool filters was not considered.
1 퐷 = 퐶 푉 퐷 Equation 6.6 푖,푗 푅 푅퐻푅푊 푖푛푔
138
Where D = dilution factor.
The volume of water ingested while swimming has been estimated as a lognormal distribution with a mean of 51.5 mL and standard deviation of 133.7 mL per swim
(lognormal parameters µ = 2.92 mL, σ = 1.43 mL) for a combined population of children and adults (Dufour et al. 2006). The swimming pool use frequency was reported by a
Netherlands study as 13-24 times per year on average (95% CI up to 65 days per year for adults and 91 days per year for children) for a six-month swimming season lasting from
May through October (Schets et al. 2011). Brisbane is coolest from June through August
(Australian Government Bureau of Meteorology 2016). Assuming that the Queensland swimming season takes place during a 9 month swimming season from September through May and that the rate of Netherlands pool swimming frequency is similar, an average and “worst case” number of swimming exposures per year of 32 and 122 were computed, respectively.
6.4.1.1.7. Toilet flushing
For toilet flushing, estimates of the volume consumed per flush range from 0.01 mL
(NRMMC-EPHC-AHMC 2006) to 0.3 mL (Schoen et al. 2014). A uniform distribution bounded by these values was used. A mean and standard deviation of 5.05 ± 2.69 flushes per day were reported (Mayer and DeOreo 1999).
6.4.1.1.8. Clothes washing
An exposure volume of 0.01 mL 100 times per year is estimated for exposure to water used for clothes washing (NRMMC-EPHC-AHMC 2006). No aerosols were observed
139 during clothes washing and therefore only ingestion is considered for clothes washing
(O'Toole et al. 2008a).
6.4.1.2. Inhalation exposure models
6.4.1.2.1. Showering, garden hosing, car washing, toilet flushing
LP concentrations are parameterized slightly differently than MAC due to the small number of detects (n = 4 positive out of 134 samples) in rainwater tanks in the study by
Hamilton et al. (2016). Maximum likelihood used for MAC is typically not appropriate for this high degree of censoring and thus a binomial method is used. Equation 6.7-6.8 are used for showering, garden hose use, car washing, and toilet flushing, accounting for the volume of aerosols of various size diameters that are large enough to hold LP or
MAC but small enough to deposit at the alveoli (1µm < diameter < 10µm). Exposure parameters for inhalation are summarized in Table 6.2.
1 1 푛 퐷표푠푒퐿푃,푗 = 퐶푅퐻푅푊,퐿푃 푃푐표푛푡푎푚퐵푡 ∑푖=1 퐶푎푒푟,푖 푉푎푒푟,i퐷퐸푖 Equation 6.7 푅 푛푠
1 퐷표푠푒 = 퐶 퐵푡 ∑푛 퐶 푉 퐷퐸 Equation 6.8 푀퐴퐶,푗 푅 푅퐻푅푊,푀퐴퐶 푖=1 푎푒푟,푖 푎푒푟,푖 푖
Where ns= number of samples, Pcontam= binomial probability of contamination with number of samples ns and probability of contamination p, Caer,i = the concentration of aerosols of diameter i where i =1:10, Vaer,i = the volume of each aerosol size bin i
140 calculated as V = (4/3)πr3, B = breathing rate (m3/min), t = exposure duration (min); and
DE = alveolar deposition efficiency of size i diameter aerosols.
The aerosol size distributions for aerosols of diameters 1-10 µm from toilet flushing, showering, and hose use are provided by O’Toole et al. (O'Toole et al. 2009,
O'Toole et al. 2008b) and are summarized in Table 6.2. For showering, a conventional showerhead operating at a water temperature of 42°C was chosen. For toilet flushing, a full 9 L flush 420 mm above the toilet was chosen; this category only observed aerosols in one size bin (median 2.5 µm). For gardening, hose use for gardening purposes is assumed to use a “conventional” nozzle on a “spray” setting while car washing would be represented by hosing with a high pressure “water efficient device” on a “jet” setting.
Experiments with the largest generation of aerosols in the 1-10 µm diameter range were chosen.
6.4.1.2.2. Pool top-up
Partitioning coefficients were used to model swimming in a pool that has been partially filled with RHRW using equations 6.9-6.10.
1 1 퐷 = 퐶 푃 푃퐵푡퐷 Equation 6.9 퐿푃,푝표표푙 푅 푅퐻푅푊,퐿푃 134 푐표푛푡푎푚
1 퐷 = 퐶 푃퐵푡퐷 Equation 6.10 푀퐴퐶,푝표표푙 푅 푅퐻푅푊,푀퐴퐶
Where P = water to air partitioning coefficient (L per m3), B = breathing rate (m3 per min), and t = exposure duration (min). Dilution D was assessed for the same conditions as the ingestion model. The partitioning coefficient for LP in an indoor warm therapy pool was calculated to range from 2.2×10-8 to 1.1 × 10-5 L per m3 in a previous study
141
(Hines et al. 2014a). A partitioning coefficient for MAC was calculated in the current study using ratios of median concentrations of nontuberculous mycobacteria in therapy pool water and air reported by Glazer et al., (2007) for sites where NTM was recovered from both air and water ranging from 1.0 × 10-4 to 5.3 × 10-3 L per m3. The number of minutes of exposure time per swim was derived from a published study (Schets et al.
2011).
6.4.2. Pathogen concentrations in RHRW
The concentrations of LP, M. avium, and M. intracellulare ranged up to 9.8 × 103 gene copies per L, 1.1 × 105 gene copies per L , and 6.8 × 105 gene copies per L, respectively
(Hamilton et al. 2016). M. intracellulare is thought to only reside in soil, while M. avium or M. chimaera are found in water (Wallace et al. 2013). The primer set used in Hamilton et al. 2016 used for M. intracellulare (Chern et al. 2015) identifies both M. intracellulare and M. chimaera. As M. intracellulare is not expected to be present in water, we assume all M. intracellulare enumerated were actually M. chimaera. Furthermore, because of the lack of a dose response model specific to M. chimaera, we assume that the dose response relationships in Hamilton et al., 2017 can be applied to the total dose of MAC (M. avium
+ M. chimaera). This presents a challenge for adding these on a sample by sample by sample basis where, for example, the concentration of M. avium was positive (above the lower limit of detection, LLOD) and above the lower limit of quantification (LLOQ), but the concentration of M. intracellulare was positive (above the LLOD) but above the
LLOQ. To address this issue, separate interval-censored lognormal distributions were fitted to each dataset (M. avium and M. intracellulare) using the package fitdistrplus in R, and the simulated distributions were added within the Monte Carlo simulation model to
142
obtain the total MAC count. Therefore, CRHRW, MAC = CRHRW,MA + CRHRW, MI (Table 6.3). It is noted that in the original study, there was no significant differences in pathogen concentrations by sampling cluster (Brisbane vs. Currumbin Ecovillage) (Hamilton et al.
2016).
6.4.3. Dose response
Exponential and Beta-Poisson dose response models are stated in equation 6.11 and 6.12, respectively. Exponential dose response model parameters for LP infection are provided in Table 6.4 (Armstrong and Haas 2007a). Exponential and Beta-Poisson dose response models for MAC pulmonary infection, disseminated infection, and cervical lymphadenitis are used (Hamilton et al. 2017). A conversion factor for pulmonary infection of 500 is applied to convert the model from the intravenous route to the inhalation route as per Hamilton et al., 2017 by dividing the daily dose by 500 within the inhalation models only. For all other models, C = 1.
−푟푑/퐶 푃푖푛푓,푑푎푖푙푦 = 1 − 푒 Equation 6.11
푑/퐶 푃 = 1 − (1 + )−훼 Equation 6.12 푖푛푓,푑푎푖푙푦 훽
6.4.4. Risk characterization
Annual risk was calculated as per Equation 6.13.
푛푓 푃푖푛푓,푎푛푛푢푎푙 = 1 − (1 − 푃푖푛푓,푑푎푖푙푦) Equation 6.13
143
Where n is the yearly frequency and f is the daily frequency of the activity. Frequency f =
1 unless otherwise stated in Table 6.1-6.2. A sensitivity analysis was conducted to identify variables contributing to uncertainty using 50,000 Monte Carlo iterations. All computations were performed in R (www.rproject.org).
In addition to annual risks for each exposure scenario, total annual infection risks for each population were calculated according to equation 6.14. A similar approach has been used to pool risks from multiple pathogens by a previous QMRA study (de Man et al.
2014c).
푃푖푛푓,푎푛푛,푡표푡푎푙,푠,푒,푝,푐 =
1 − (1 − 푋푎1푃푖푛푓,푎푛푛,푎1)(1 − 푋푎2푃푖푛푓,푎푛푛,푎2 ) … (1 − 푋푎푛푃푖푛푓,푎푛푛,푎푛) Equation 6.14
Where Pinf,ann,total,s,e,p,l = the total annual risk incurred from each scenario s where s = inhalation or ingestion annual risk from activities a1, a2, a3..an where an = showering, drinking, etc., and Xa,n = portion of sampling cluster location c that uses rainwater for each activity (Appendix Table 9.20). Total risk is calculated for health endpoint e = pulmonary infection, cervical lymphadenitis, or disseminated infection and p = healthy populations, children, or vulnerable/immune-compromised populations that are relevant for each of the previous endpoints, respectively. For the pulmonary infection endpoint,
LP and MAC risks are included in the same equation to compute total pulmonary disease risks, assuming the infection risks from these organisms are additive and that infection with one pathogen would not predominate over or preclude infection with the other.
144
The Spearman rank correlation coefficient was used to identify the most important predictive factors of annual infection or clinical severity infection risk, were 0 is no influence and -1 or +1 when the output is wholly dependent on that input. The model inputs were ranked based on their correlation coefficient with the output variable, annual risk.
145
Table 6.1 Monte carlo simulation input parameters for ingestion scenarios (for MAC only)
Parameter Symbol Unit Value Distribution Source Drinking water a Intake rate Ving,dw L per day µ = -1.329, σ = 1.542 Lognormal (USEPA 2011a) Exposure frequency ndw - 365 Point Assumption Log removals due to filtration L logs Min=0.4, Max=2 Uniform (Jordan 2008) Toilet flushing Volume ingested Ving,t mL Min = 0.01, Max = 0.3 Uniform (NRMMC-EPHC-AHMC 2006, Schoen et al. 2014) Toilet flushes per day nt Flushes per day µ =1.494, σ = 0.500 Lognormal (Mayer and DeOreo 1999) Showering Volume ingested Ving,sh mL Min = 0.058, Max = Uniform (Ahmed et al. 2010) 1.9 Showers per year nsh Number per year 365 Point Assumption Garden hosing -9 Volume ingested Ving,gh mL Min = 2×10 , Max = Uniform (Ahmed et al. 2010, 1.1×10-3 NRMMC-EPHC-AHMC 2006, Schoen et al. 2014) Hosing events per year ngh Number per year 90 Point (NRMMC-EPHC-AHMC 2006) Car washing Volume ingested Ving,cw mL Min = 0.06, Max = Uniform (Sinclair et al. 2016) 3.79 Car washing events per year ncw Number per year 12 Point (Villarreal and Dixon 2005) Pool top-up Dilution Factor D % 10 Point Assumption 50 90 Volume ingested per swim Ving,sw mL µ = 2.92 , σ = 1.43 Lognormal (Dufour et al. 2006) Swims per year nsw Number per year Average 32 Point Assumptions based on Worst case 122 (Schets et al. 2011) Clothes washing Volume ingested Ving,cw mL 0.01 Point (NRMMC-EPHC-AHMC 2006) Number of times clothes are ncw Number per year 100 Point (NRMMC-EPHC-AHMC washed with rainwater per year 2006) Produce consumption
Volume irrigated water retained VR mL per g µ = 0.108 , σ = 0.019 Normal, truncated (Shuval et al. 1997) on lettuce at 0
146
-5 Internalized fraction of MAC in Fint Proportion Min=1.13 × 10 , Uniform (Kaevska et al. 2014) irrigation water Max=9.49 × 10-4 In-field decay on surface of plant Kf,s d-1 -0.0484 Point (Cook et al. 2013) Log reductions due to lettuce W Logs µ = 0.3, σ = 0.1 Normal, truncated (Holvoet et al. 2014) washing at 0 Time in field between irrigation tf day 2 Point (Barker et al. 2013) and harvest Intake rate I g per person per 40 Point (NRMMC-EPHC-AHMC day 2006) Lettuce consumption events per n Days per year 70 Point (NRMMC-EPHC-AHMC year 2006) aLognormal parameters mean, standard deviation (µ, δ) calculated from population (normal) parameters (푥̅, s) using standard formulae as follows: µ = ln(푥̅2/(s2+푥̅2)1/2), δ = [ln(1+( s2/푥̅2))]1/2, where 푥̅ is the sample mean and s2 is the sample standard deviation.
147
Table 6.2 Monte carlo simulation input parameters for inhalation scenarios
Parameter Symbol Unit Value Distribution Source Breathing rate, light activity, B m3 per min 0.013-0.017 Uniform (USEPA 2011a) breathing cycle period 8 s and 1 L tidal volume Deposition efficiency (diameter) DEi Fraction Uniform (Heyder et al. 1986) 1 Min = 0.23, Max =0.25 2 Min = 0.4, Max =0.53 3 Min = 0.36, Max =0.62 4 Min = 0.29, Max =0.61 5 Min = 0.19, Max =0.52 6 Min = 0.1, Max =0.4 7 Min = 0.06, Max =0.29 8 Min = 0.03, Max =0.19 9 Min = 0.01, Max =0.12 10 Min = 0.01, Max =0.06 Toilet flushing a Toilet flushes per day nt Flushes per day µ =1.494, σ = 0.500 Lognormal (Mayer and DeOreo 1999) Time in bathroom after flush tt Min per flush Min = 1, Max = 5 Uniform Lim et al. (2015) 3 Concentration of aerosols of Caer,i # aerosols per cm µ = -1.246, σ = 1.885 Lognormal (O'Toole et al. 2009) diameter i: of air 2.5
Showering Shower duration tsh min per day 15 Point (Schoen and Ashbolt 2011) Showers per year fsh Showers per year 365 Point Assumption 3 Concentration of aerosols of Caer,i # aerosols per cm Lognormal (O'Toole et al. 2009) diameter i: of air 1.5 µ = 3.718, σ = 0.296 2.5 µ = 3.699, σ = 0.170 4.5 µ = 5.549, σ = 0.348 8 µ = 6.185, σ = 0.309 Garden hosing Hosing duration tgh min 7 Point (Ahmed et al. 2010) Hosing events per year ngh Number per year 90 Point (NRMMC-EPHC-AHMC 2006) 3 Concentration of aerosols of Caer,i # aerosols per cm Lognormal (O'Toole et al. 2009) diameter i: of air 1.5 µ = 5.728, σ = 0.274
148
2.5 µ = 4.949, σ = 0.333 4.5 µ = 3.047, σ = 0.586 8 µ = -2.451=, σ=1.579 Car washing Car washing duration tcw min 10 Point (O'Toole et al. 2008b) Car washing events per year ncw Number per year 12 Point (Villarreal and Dixon 2005) 3 Concentration of aerosols of Caer,i # aerosols per cm Lognormal (O'Toole et al. 2009) diameter i: of air 1.5 µ = 6.187, σ = 0.476 2.5 µ = 4.665, σ = 0.420 4.5 µ = 1.742, σ = 0.591 8 µ = -1.551, σ = 0.833 Pool top-up 3 -8 Partitioning coefficient- L. PCLP L per m Min = 2.2×10 , Max = Uniform (Hines et al. 2014a) pneumophila 1.1 × 10-5 3 -4 Partitioning coefficient- MAC PCMAC L per m Min = 1.0 × 10 , Max Uniform (Glazer et al. 2007) = 5.3 × 10-3 Dilution Factor D % 10 Point Assumption 50 90 Pool exposure time per swim tpool min µ = 4.2, σ = 0.55 Lognormal (Sales-Ortells and Medema 2012, Schets et al. 2011) Swims per year nsw Number per year Average 32 Point Assumptions based on Worst case 122 (Schets et al. 2011) aLognormal parameters mean, standard deviation (µ, δ) calculated from population (normal) parameters (푥̅, s) using standard formulae as follows: µ = ln(푥̅2/(s2+푥̅2)1/2), δ = [ln(1+( s2/푥̅2))]1/2, where 푥̅ is the sample mean and s2 is the sample standard deviation.
149
Table 6.3 Monte carlo simulation input parameters for pathogen concentrations
CONCENTRATIONS IN RHRW a M. avium CRHRW,MA # per L µ = 0.723, σ = 4.349 Lognormal (Hamilton et al. 2016) M. intracellulare CRHRW,MI # per L µ = 6.720, σ = 2.410 Lognormal (Hamilton et al. 2016) L. pneumophila, positive samples CRHRW,LP # per L µ = 8.080, µ = 0.745 Lognormal (Hamilton et al. 2016) only Probability of L. pneumophila Pcontam Fraction n = 134, p = 0.03 Binomial (Hamilton et al. 2016) occurrence Recovery efficiency R Fraction µ = 0.84, σ = 0.32 Normal, truncated (Hamilton et al. 2016) at 0 aLognormal parameters mean, standard deviation (µ, δ) calculated from population (normal) parameters (푥̅, s) using standard formulae as follows: µ = ln(푥̅2/(s2+푥̅2)1/2), δ = [ln(1+( s2/푥̅2))]1/2, where 푥̅ is the sample mean and s2 is the sample standard deviation.
150
Table 6.4 Monte carlo simulation dose response input parameters
Parameter Symbol Unit Value Distribution Source L. pneumophila Dose response parameter for L. r Unitless µ = -2.934, σ = 0.488 Lognormala (Armstrong and Haas pneumophila, infection endpoint 2007a) MAC Dose response parameter for M. r Unitless µ = -13.742, σ = 0.208 Lognormal (Hamilton et al. 2017) avium- pulmonary infection (subclinical) Conversion factor from C Unitless 500 Point (Hamilton et al. 2017) intravenous to inhalation route for pulmonary infection model Dose response parameters for α Unitless 0.201 Point (Hamilton et al. 2017) disseminated infection β 1.15 × 10-6 Dose response models for cervical r Unitless µ = -17.714, σ = 1.648 Lognormal (Hamilton et al. 2017) lymphadenitis in children aLognormal parameters mean, standard deviation (µ, δ) calculated from population (normal) parameters (푥̅, s) using standard formulae as follows: µ = ln(푥̅2/(s2+푥̅2)1/2), δ = [ln(1+( s2/푥̅2))]1/2, where 푥̅ is the sample mean and s2 is the sample standard deviation.
151
6.5. Results
6.5.1. Ingestion
Annual risks for each ingestion scenario are shown in Figure 6.2 and compared to a hypothetical 1 × 10-4 drinking water annual infection risk benchmark (Regli et al. 1991).
Consumption of rainwater presented the highest risk for both cervical lymphadenitis in children and disseminated infection for vulnerable/ immune-deficient adults. In all drinking water scenarios, the median annual risk was > 1 × 10-4. Consumption of lettuce irrigated with reclaimed water, showering, use of a garden hose, car washing, and toilet flushing had median annual risks below the benchmark, but the 95% confidence limit exceeded this value in all cases except for children via the car washing route. There is no risk benchmark for swimming in pools, although there are recreational standards for microbiological risks associated with freshwater and marine beaches of 8 and 19 cases of highly credible gastrointestinal illness (HCGI) per 1,000 recreators in fresh and marine waters, respectively (USEPA 2011b). Eight cases of HCGI per 1,000 is equivalent to 36 cases of gastrointestinal illness per 1,000 recreators based on a more encompassing definition of gastrointestinal illness used in the USEPA National Epidemiological and
Environmental assessment of Recreational Water (NEEAR) studies (USEPA 2011b).
These definitions are not particularly compatible with the health endpoints used in this analysis, and are therefore not used for comparison here. Generally, swimming pool risks were higher for vulnerable/immune-deficient populations than for children. Swimming pool water would need to be diluted to 90% sterile water: 10% RHRW in order to have a median risk below 1 × 10-4.
152
A sensitivity analysis for ingestion risk scenarios is shown in Table 6.5. The most influential predictor of variability in annual risk for all scenarios was the concentration of
MAC in RHRW (CRHRW) (Spearman rank correlation coefficients ranging from 0.66-
0.98). However, as CRHRW was calculated by adding the concentration of M. avium and M. intracellulare in RHRW, the sensitivity analysis determined that the concentration of M. intracellulare was more influential as this is more commonly present in RHRW and in higher concentrations than M. avium on average. The dose response parameter r and volume of water ingested (Ving) or retained by lettuce (VR) were the second and third most influential parameters, respectively.
153
Figure 6.2 Annual cervical lymphadenitis (children) or disseminated infection (vulnerable/immune compromised populations) risks for ingestion of Mycobacterium avium complex through food or water exposure scenarios. Median and 95% confidence intervals and scenario analysis for various pool dilution levels (D = 10%, 50%, or 90% RHRW: sterile water), with or without drinking water under-sink point of use filtration (filtration/no filtration), and swimming frequency assumptions (32 or 122 swims per year) are shown
154
6.5.2. Inhalation
Annual risks for each inhalation scenario are shown in Figure 6.3. The median annual LP risks for showering, car-washing, and all pool top-up scenarios exceeded a 1 × 10-4 benchmark. Risks were highest for shower and pool top-up. For MAC, all 95% confidence intervals were below the benchmark value.
A sensitivity analysis for inhalation risk scenarios is shown in Table 6.6. The most influential predictor of variability in annual risk for toilet flushing, showering, garden hosing, or car washing was the concentration of LP and/or MAC in RHRW (CRHRW)
(Spearman rank correlation coefficients ranging from 0.59- 0.97). For LP toilet flushing scenario only, the concentration of aerosols (Caer) was the most important predictor (ρ =
0.78) while for MAC toilet flushing, Caer was the second most important predictor (ρ =
0.56). For LP showering and garden hosing scenarios, the probability of contamination
(Pcontam) was also an important factor (Spearman rank correlation coefficients ranging from 0.45-0.46). For all pool scenarios, the partitioning coefficient (P) was either the first or second-most important predictor (Spearman rank correlation coefficients ranging from
0.29-0.50).
155
Figure 6.3 Annual infection risks for inhalation of Mycobacterium avium complex or L. pneumophila for various exposure scenarios. Median and 95% confidence intervals and scenario analysis for various pool dilution levels (D = 10%, 50%, or 90% RHRW: sterile water) and swimming assumptions (32 or 122 swims per year) are shown
156
Table 6.5 Sensitivity analysis with Spearman rank correlation coefficients for ingestion risk scenarios. Risks for endpoints of children with cervical lymphadenitis and severe immune deficiency with disseminated infection are shown
Scenario Population Parameter (Spearman rank correlation coefficient)
CRHRW CMA CMI n Fint L r R Ving W Drinking water Children 0.659 0.144 0.599 -0.304 0.456 -0.119 0.427 Vulnerable 0.749 0.158 0.678 -0.336 -0.135 0.498 Drinking water- no filtration Children 0.684 0.145 0.622 0.489 -0.127 0.460 Vulnerable 0.808 0.163 0.735 -0.148 0.524 Lettuce Children 0.785 0.162 0.716 0.002 0.554 0.069* -0.087 Vulnerable 0.970 0.190 0.888 -0.001 0.065* -0.091 Shower Children 0.760 0.161 0.691 0.531 -0.134 0.240 Vulnerable 0.923 0.187 0.844 -0.158 0.285 Hose Children 0.749 0.153 0.684 0.526 -0.141 0.272 Vulnerable 0.901 0.172 0.823 -0.162 0.323 Car Children 0.758 0.154 0.690 0.530 -0.144 0.252 Vulnerable 0.915 0.182 0.832 -0.162 0.309 Toilet Children 0.747 0.142 0.682 0.166 0.527 -0.144 0.226 Vulnerable 0.906 0.176 0.828 0.190 -0.165 0.279 Clothes washing Children 0.793 0.169 0.719 0.549 -0.143 Vulnerable 0.978 0.195 0.892 -0.172 Pool, D = 10%, n = 32 swims Children 0.704 0.145 0.640 0.494 -0.127 0.431 Vulnerable 0.827 0.170 0.753 -0.153 0.498 Pool, D = 50%, n = 32 swims Children 0.700 0.130 0.639 0.486 -0.130 0.435 Vulnerable 0.827 0.162 0.753 -0.151 0.505 Pool, D = 90%, n = 32 swims Children 0.701 0.142 0.637 0.496 -0.133 0.423 Vulnerable 0.827 0.168 0.755 -0.152 0.497 Pool, D = 10%, n = 122 swims Children 0.701 0.144 0.638 0.493 -0.130 0.432 Vulnerable 0.825 0.160 0.751 -0.142 0.493 Pool, n = 50%, n = 122 swims Children 0.703 0.145 0.640 0.490 -0.128 Vulnerable 0.824 0.169 0.750 -0.159 0.493
157
Pool, D = 90%, n = 122 swims Children 0.703 0.144 0.639 0.493 -0.131 -0.131 Vulnerable 0.826 0.165 0.754 -0.146 0.486
*For the lettuce scenario, the coefficient for volume of water retained by lettuce (VR) is shown
158
Table 6.6 Sensitivity analysis with Spearman rank correlation coefficients for inhalation risk scenarios. Risks for endpoints of pulmonary infection in healthy populations with L. pneumophila or MAC are shown.
Scenario Path-ogen Parameter (Spearman rank correlation coefficient)
B Caer Caer1.5 Caer2.5 Caer4.5 Caer8 CRHRW CMA CMI DE DE1.5 DE2.5 DE4.5 DE8 f P Pcontam r R t Toilet flushing LP 0.03 0.78 0.30 0.05 0.21 0.26 0.20 -0.17 0.18 MAC 0.02 0.56 0.70 0.14 0.64 0.03 0.10 0.06 -0.13 0.13 Shower LP 0.06 -0.004 0.001 0.09 0.16 0.59 0.01 -0.01 0.04 0.18 0.46 0.38 -0.33 MAC 0.03 -0.01 0.001 0.05 0.08 0.96 0.19 0.88 -0.002 -0.002 0.03 0.09 0.08 -0.17 Garden hose LP 0.07 0.04 0.13 0.17 0.01 0.60 0.03 0.06 0.04 0.00 0.46 0.39 -0.34 MAC 0.03 0.02 0.06 0.08 0.01 0.97 0.19 0.88 0.02 0.02 0.03 0.01 0.09 -0.18 Car wash LP 0.08 0.14 0.17 0.07 0.09 0.70 0.07 0.05 0.02 0.04 0.00 0.45 -0.39 MAC 0.04 0.06 0.08 0.02 0.05 0.97 0.19 0.88 0.03 0.03 0.01 0.01 0.09 -0.17 Pool, D=10%, n=32 LP 0.05 0.46 0.50 0.37 0.29 -0.26 0.33 MAC 0.02 0.89 0.17 0.82 0.29 0.08 -0.16 0.20 Pool, D=50%, n=32 LP 0.04 0.45 0.50 0.37 0.29 -0.26 0.34 MAC 0.03 0.89 0.18 0.82 0.29 0.08 -0.16 0.21 Pool, D=90%, n=32 LP 0.05 0.45 0.50 0.37 0.30 -0.26 0.33 MAC 0.03 0.89 0.18 0.81 0.28 0.09 -0.17 0.20 Pool, D=10%, n=122 LP 0.04 0.45 0.50 0.36 0.29 -0.27 0.33 MAC 0.03 0.89 0.18 0.81 0.29 0.08 -0.16 0.20 Pool, D=50%, n=122 LP 0.04 0.44 0.50 0.37 0.29 -0.26 0.33 MAC 0.03 0.89 0.18 0.81 0.30 0.08 -0.16 0.20 Pool, D=90%, n=122 LP 0.04 0.45 0.50 0.35 0.30 -0.26 0.34 MAC 0.03 0.89 0.18 0.81 0.29 0.08 -0.16 0.21
159
6.5.3. Total annual risks
In order to quantify the risks for each relevant population/ exposure route, a distribution of total annual risks was computed for each scenario. For cervical lymphadenitis in children and disseminated infection in vulnerable/immune-deficient populations via a combination of all oral ingestion exposure routes (Figure 6.4), risks for vulnerable populations were higher than for children, and higher for the
Currumbin Ecovillage study population than the Brisbane study population based on survey responses. Annual risks were not substantially impacted by the dilution of
RHRW used in pools with sterile water, the number of swims per year, or the use of drinking water POU filters (<1 log). The use of POU filters had the least impact on total risk of these simulated scenarios. All median vulnerable population total risks were above a 1 × 10-4 benchmark while total risks for children were below this benchmark for all scenarios.
Total inhalation risks from all aerosol-generating activities reported in the survey for each sampling cluster are summarized in Figure 6.5. Median total annual LP pulmonary infection risks were approximately 5 orders of magnitude higher than
MAC pulmonary infection risks for all scenarios, and all 95% confidence intervals for
LP total risks were above the 1 × 10-4 benchmark while all those for MAC were below this value. Therefore, total pulmonary infection risks (LP + MAC) were driven by LP risks rather than MAC risks. Total annual risks for LP were slightly higher for the Brisbane compared to the Currumbin Ecovillage, while total annual MAC risks were lower for Currumbin than for Brisbane. The number of swims per year and pool dilution had only a small impact (<1 log) on differences in risk between scenarios. 160
Figure 6.4 Total annual risks from all activities for cervical lymphadenitis in children or disseminated infection in vulnerable/immune compromised populations via ingestion. Median and 95% confidence intervals shown. Scenario analysis for various pool dilution levels (D = 10%, 50%, or 90% RHRW: sterile water), with or without drinking water under-sink point of use filtration (filtration/no filtration), and swimming frequency assumptions (32 or 122 swims per year) are shown.
161
Figure 6.5 Total annual pulmonary infection risks from all activities via inhalation of Mycobacterium avium complex (MAC), L. pneumophila (LP), or both organisms (LP+MAC). Median and 95% confidence intervals are shown. Scenario analysis for various pool dilution levels (D = 10%, 50%, or 90% RHRW: sterile water) and swimming assumptions (32 or 122 swims per year) are shown.
162
6.6. Discussion
The current study is the first investigation to assess MAC risks for human- relevant, non-MAP species of MAC in an aquatic environment. For MAC, the drinking water route of exposure presents the greatest risks for susceptible populations including children and the immune-deficient. For susceptible populations, it would not be appropriate to drink rainwater. The use of a filter did not have a substantial impact on simulated annual risks in the scenario analysis; additionally, opportunistic pathogens such as MAC can colonize POU filtration systems (Rodgers et al. 1999), limiting their utility for mitigating drinking water risks.
For swimming pool use, risks can be lowered through dilution with treated water.
This work indicates that a dilution to 10% rainwater would bring median risks below a drinking water benchmark, which may be more appropriate for comparison with susceptible populations than healthy populations in order to reduce risks. However, this dilution factor can be lessened depending on the recreational water quality criteria used for comparison. Legionella pneumophila risks were above the drinking water benchmark for all scenarios except toilet flushing and were greater than MAC risks in all cases. Showering and swimming pool use presented the highest risks for pulmonary infection.
Both inhalation and ingestion simulations indicate that toilet flushing is one of the lowest-risk scenarios and is likely to be an appropriate use for rainwater. All individual MAC pulmonary infection inhalation risks for healthy populations were below the drinking water benchmark. This could be affected by the fact that a conversion factor of 500 was used to convert the intravenous to inhalation dose response route (Hamilton et al. 2017). Due to the lack of dose response models for
163 pulmonary infections in susceptible populations, further investigation of potential dose response models for this purpose is recommended.
An important factor in this work was the regional preference for various uses of
RHRW. When examining differences between two sampling clusters in Brisbane and the Currumbin Ecovillage, ingestion risks and inhalation risks for total inhalation pathogen load (LP + MAC) were higher for the Currumbin Ecovillage compared to
Brisbane. This is likely due to the higher degree of RHRW application for potable uses in Currumbin compared to Brisbane. When examining total inhalation risks from all uses (showering, garden hose, car washing, etc.) and considering L. pneumophila and MAC risks separately, Currumbin risks for LP were higher than Brisbane but
MAC risks were lower than Brisbane. This may be due to higher pool and shower usage in Brisbane. MAC pool inhalation risks were computed using a partitioning coefficient, which is reported to be higher for MAC than for LP. LP shower risks were the highest compared to other scenarios, indicating that showering may have had a greater influence on total annual pulmonary risk for Currumbin compared to the other scenarios.
For previous studies focusing on LP risks in RHRW with LP concentrations of 60-
170 gene copies (assumed equivalent to cells) per L rainwater, between 3.0 × 10-2 and
8.6 × 10-2 LP infections per 10,000 exposed people per shower exposure and 1.8 × 10-
2 – 5.1 × 10-2 infections per 10,000 exposed people per hosing exposure were expected (Ahmed et al. 2010). This would be between an annual probability of infection of approximately 1.09 × 10-3 and 3.13 × 10-3 and is comparable with the risks calculated in the current study. A case of Legionnaires disease has also been previously linked to the usage of a garden hose (Piso et al. 2007).
164
Another study of rainwater used as a source of water for splash parks estimated a mean Legionella infection risk of 9.3 × 10-5 for a 3.5 minute splash park exposure for children and 1.1 × 10-4 for adults, but that for a 2 h exposure this risk would be approximately 2.8 × 10-3 (de Man et al. 2014a). Although the pool exposure model is different than that used in the current study, de Man et al. work supports that pool risks can be potentially high compared to other exposure routes. The current QMRA did not consider ingestion of mouthfuls of water as in de Man et al. models due to the difference between active splash parks and private pools. However, exposures to pool water could potentially be higher in some circumstances if mouthfuls of water are considered. Private household pool or whirlpool use has indeed been linked to
Legionnaires disease cases (Euser et al. 2010). It is noted in the current QMRA that although dilution of RHRW with sterile water was considered, RHRW users may dilute rainwater with treated tap-water which may not be free of LP and MAC but is likely to contain fewer pathogens than RHRW (Whiley et al. 2014a). Additionally, some chlorine residual present in municipal drinking water could contribute to pathogen removal. However, LP (Cooper and Hanlon 2010) , and especially MAC
(Falkinham III 2003, Steed and Falkinham III 2006, Taylor et al. 2000) are resistant to disinfection including chlorination, and would not be likely to decline significantly at low levels of chlorine residual. Chlorine treatment in pools by owners could contribute to pathogen removal and could be considered in a more detailed assessment of opportunistic pathogens in pools.
Similarly to the previous QMRAs for RHRW (Ahmed et al. 2010, de Man et al.
2014a), in the absence of information regarding the relationship between total gene copies and the viability of pathogens in rainwater tanks, one gene copy was considered equivalent to one viable, infectious pathogen. This assumption could
165 potentially lead to an overestimation of risks; however, culture-based assessments of pathogens can neglect to quantify viable but non-culturable (VBNC) microorganisms and therefore underestimate risks. If a 1% or 10% viability assumption is made, total annual risks would be 1 – 2 orders of magnitude lower, respectively.
Finally, regarding the clothes washing scenario, MAC can be present in fecal material (Yajko et al. 1993) that can be introduced to the laundry cycle and provide additional pathogen loading. The focus of this QMRA was on the risks of RHRW, and therefore this additional input was not considered. A more in-depth QMRA model for
MAC risks due to clothes washing might consider transfer of fecal-associated MAC to water and hand-to-mouth contact after clothes laundering and impact of any disinfectants used (Callewaert et al. 2015, Gerba and Kennedy 2007, Gibson et al.
1999, Lopez et al. 2013).
Regarding the management of RHRW risks, toilet lids can be closed prior to flushing to prevent the spread of mists. Similarly, less aerosol-efficient showerheads can be installed in showers for which rainwater is used as a source. For pool-users, pool filters can be installed and rainwater can be diluted with treated water in order to potentially lower risks. In this study, only under-sink POU filters were considered for mitigating drinking water risks, however, other methods such as solar-disinfection have been explored for use with rainwater tanks which could provide higher log removals (Reyneke et al. 2016). Treatment systems designed to treat all water entering the home from a rainwater tank as opposed to only water obtained via a kitchen sink tap would be beneficial for reducing risks. More exploration of in situ field performance of various rainwater treatment systems over time is warranted as a means for approaching pathogen risk mitigation for RHRW.
166
6.7. Conclusions
Regional use preferences for roof-harvested rainwater had an impact on the
rank order of ingestion and inhalation risks for opportunistic pathogens
Based on the results of the quantitative microbial risk assessment, harvested
rainwater is not recommended for drinking in child or immune-compromised
populations
For healthy populations, showering and car-washing with high pressure sprays
are not recommended for harvested rainwater
Compared to Mycobacterium avium infection risks, potential Legionella risks
should drive risk mitigation strategies for the inhalation route of exposure
6.8. Acknowledgements
This work was supported by a Fulbright-CSIRO Postgraduate scholarship sponsored by the Australian-American Fulbright Commission. The authors are grateful to rainwater study participants for providing survey information used in this work and for CSIRO staff Dr. Jatinder Sidhu, Leonie Hodgers, Andrew Palmer, Kylie Smith, and Pradip Gyawali for their contributions to the original pathogen quantification study used in this risk assessment.
167
7. Quantitative microbial risk assessment (QMRA) of Legionella in reclaimed water5
7.5. Introduction
Growing global water scarcity has intensified the need to recover water resources from wastewater, especially as population growth, economic development, and urbanization increase pressures on existing water supplies (Levine and Asano
2004). Reclaimed water can alleviate stress on municipal water systems and augment existing water portfolios. However, there is the potential for health risks from human contact with contaminants in reclaimed water through inhalation or ingestion of water sprays.
Agricultural and industrial water reuse represent the sectors with the largest reclaimed water usage in the United States (Jiménez and Asano 2008). Reclaimed water for cooling system purposes further represents the largest industrial water reuse application (Metcalf & Eddy 2007). Cooling systems may consume 20-50% of a facility’s water usage (Aoki et al. 2005). Common uses of reclaimed water such as spray irrigation or cooling towers can produce aerosols that are of concern because bacteria such as Legionella pneumophila can travel beyond the immediate vicinity of application (Li et al. 2011). To inform appropriate usages of reclaimed water and identify factors which have the greatest implication for best management practices, a quantitative microbial risk assessment (QMRA) is presented for scenarios of toilet flushing, spray irrigation, and cooling tower generated aerosols. Quantitative
Microbial Risk Assessment is a framework that integrates information regarding pathogen occurrence, infectivity, and exposure for determining the health implications
5 This chapter is in press: LeChevallier, M.W., Bukhari, Z., Jjemba, P., Johnson, W., Haas, C.N. and Hamilton, K.A. (2017) Development of a risk management strategy for Legionella in recycled water systems (WRF12-05), WateReuse Research Foundation, Alexandria, VA. Mark T. Hamilton is acknowledged for his contributions to model code for this chapter. 168 of microbial hazards. The QMRA is conducted using a process of hazard identification, exposure assessment, dose response assessment, and risk characterization (Haas et al. 2014). General approaches for Legionella QMRA utilized in this work are summarized in Section 9.6 (see Appendix Figure 9.21- Figure 9.22).
7.6. Hazard identification
Legionella is an opportunistic waterborne pathogen of significant public health importance known to occur in engineered water systems, ambient water environments, and soils. Legionella spp. grow in biofilms in piping and can slough off and become aerosolized through water fixtures, at which point human exposure can occur. To date, no infectious disease outbreaks have been reported in association with reclaimed water. However, concerns exist regarding aerosol production of Legionella as outbreaks have been associated with exposure to aerosols generated by cooling towers (Castilla et al. 2008b, Nguyen et al. 2006, Walser et al. 2014), decorative fountains (Haupt et al. 2012), and other common water uses (Hines et al. 2014b). A recent exposure study of irrigation workers exposed to reclaimed water showed higher colonization with Staphylococcus and Enterococcus bacteria compared to office workers, although this difference was not significant (Goldstein et al. 2014). Few studies have quantified Legionella in reclaimed water (Jjemba et al. 2010, Palmer et al. 1995), and a QMRA has not yet been performed for this pathogen for exposures to reclaimed water.
7.7. Methodology
7.7.6. Exposure Models
169
Three exposure scenarios of toilet flushing, spray irrigation, and cooling tower-generated aerosols were considered. All models followed previously reviewed exposure model steps summarized in Appendix Figure 9.9. Generally, Legionella was considered to be ubiquitously present in reclaimed water, with aerosols generated at rates specific to each process modeled. Not all aerosols released from a given activity will reach a receptor while possessing droplet diameters within the respirable range.
Aerosol particles of diameters between 1 and 10 µm were considered respirable for all three exposure scenarios, as Legionella is typically 1 to 2 µm long and 0.3 to 0.9 µm wide, and particles greater than 10 µm are not likely to reach the respiratory tract
(Baron and Willeke 1986, Metcalf & Eddy 2007). Where available, information was used regarding how the bacteria partitions in aerosol, and pathogen decay during transport through air to a receptor. Most large aerosol droplets are trapped in the nasopharyngeal region, and smaller particles are able to travel to the alveoli, where
Legionella infection is initiated and behaves according to an exponential dose response model. Human exposure patterns for each scenario were taken into account to annualize risks. Where necessary, data from published graphs necessary for aerosol calculations were extracted using Digitize It® v. 4.0.2 (Alcasa, 2010).
7.7.6.1. Toilet flushing
Three QMRA methods were compared to assess infection risks from toilet flushing. Method 1 for toilet flushing is modified from the QMRA exposure model of
Schoen and Ashbolt (2011) originally developed for showering as per Equation 7.1 using an upper bound partitioning coefficient (PC) derived by Hines et al. (2014b)
(1.3× 10-6 L/m3) and lower bound calculated from a controlled toilet flushing experiment using gram negative bacteria Serratia marcesens NCTC 10211 (Barker and Jones 2005). Barker and Jones (2005) seeded toilets with 1010 bacteria on the
170 toilet sidewalls and flushed 5 minutes after applying the inoculum. The bulk toilet water contained 108 CFU/ mL Serratia prior to flushing. Bacterial air samples were collected 20 cm above and 30 cm in front of the toilet after flushing and a maximum concentration of 1370 CFU/ m3 (SE 527) in air was detected 1 minute after the flush resulting in a PC of 1.37× 10-8 L/m3. This is lower than previous PC values used in
Legionella risk assessments and related work: 1) Showering PC based on
Brevundimonas diminuta of 5.l8× 10-6-1.64× 10-5 CFU m-3/ L-3 (Schoen and Ashbolt
2011); 2) Hot springs aerosol PC 2.3× 10-5 CFU m-3/ CFU L-1 based on endotoxin data
(Armstrong and Haas 2008) ; 3) Bursting bubbles in distilled water at 22°C seeded with Serratia marcescens with PC 1× 10-6 CFU m-3/CFU L-1 (Blanchard and Syzdek
1982); and whirlpool spa PC ranging from <3× 10-6 (no air injection)-1.1× 10-3 L/m3
(air injection) based on Pseudomonas aeruginosa, which was deemed with the experiment to be a more appropriate surrogate than MS-2 coliphage tested in the study due to its size and similarities with Legionella (Moore et al. 2015). A lower generation rate of bacteria-containing aerosol is expected for toilet flushing due to the less active generation process than for showering or aerated or hot spring spas. It is assumed the decay in aerosol for toilet flushing is negligible over a 1-5 minute exposure event directly after flushing.
The fraction (Fi) of Legionella that partitions into various aerosol sizes after transitioning from bulk water to aerosol is described by Schoen and Ashbolt (2011) as
0.75, 0.09, and 0.14 for aerosols of 1-4µm, 5-6µm, and 7-10µm, respectively.
Deposition efficiencies (DE) derived by Schoen and Ashbolt (2011) from Schlesinger
(1985) for these size bins are 0.2-0.54, 0.1-0.65, and 0.01-0.1, respectively.
푛 퐷표푠푒퐿푒푔,푝푓 = 퐶퐿푒푔푃퐶푤푎퐼푡 ∑푖=1 퐹푖퐷퐸푖 Equation 7.1
171
Where DoseLeg,pf= the dose of Legionella deposited in the lungs per toilet flush;
CLeg,di= the concentration of Legionella in bulk water; PCwa= bacterial water to air partitioning coefficient; I= the mean inhalation rate of air breathed after toilet flushing
(m3 air/ min); t is the exposure duration or time spent in the room after toilet flushing
(min); F= the fraction of Legionella that partitions to a given size range i; DE= the alveolar deposition efficiencies of aerosols in each size range bin i; and i is represented by bins of 1-5, 5-6, and 6-10 (Schoen and Ashbolt 2011).
Method 2 is modified from the approach of Lim et al. (2015) using the concentration of 2.5 µm median diameter aerosols produced by a toilet flush measured 420 mm above a toilet by a cistern toilet suite O'Toole et al. (2009)
(Equation 7.2). The model was a Caroma Uniset cistern model P/N 213012 and pan model P/N 601200W, operated at full capacity for either 9L/4.5 L for full/half flush.
Aerosols ranging from 0.06 to 20 µm were measured, however, aerosols were only observed in the 2-3 µm size bin and none of the other 3-10 µm bins for a single toilet flush. No aerosols were observable for a half-flush. The experiment also evaluated a continuous flushing scenario which observed aerosol droplets over the 3-6 and 6-10
µm ranges, however this was not selected due to the implausibility of a continuous flush exposure scenario. The deposition efficiency for aerosols of 2.5 µm reported by
Lim et al. (2015) derived from Heyder et al. (1986) is used.
퐷표푠푒퐿푒푔,푝푓 = 퐶퐿푒푔퐶푎푒푟표,2.5푉푎푒푟표,2.5퐼푡퐷퐸2.5 Equation 7.2
Where DoseLeg,pf= the dose of Legionella deposited in the lungs per toilet flush; CLeg= the concentration of Legionella in bulk water; Caero,2.5= the concentration of aerosols
172
of diameter 2.5 µm measured 420 mm above the toilet by O'Toole et al. (2009); Vaero,
3 - 2.5= the volume of 2.5 µm aerosols calculated as V=(4/3)πr where d = 2r = 2.5×10
6m; I= the mean inhalation rate of air breathed after toilet flushing (m3 air / min); t is the exposure duration or time spent in the room after toilet flushing (min); and DE= the alveolar deposition efficiency of 2.5 µm aerosols.
Method 3 uses the approach of Lim et al. (2015) as above, but uses recent data on the aerosol generation rate from modern flush toilets provided by Johnson et al.
(2013) (Equation 7.3). Four types of toilets were characterized including a pre-FEPA gravity flow toilet (13.3 Lpf), a dual-flush high-efficiency toilet (HET) (3.8 or 4.9
Lpf), a dual-flush pressure-assisted gravity flow toilet (PAT) (4.2 or 4.9 Lpf), and a flushometer (FOM) toilet) (5.3 Lpf). Data were available for 1-10 µm size bins and were digitally extracted and averaged from each toilet type from Johnson et al. (2013)
(Figure 6 in the Johnson paper), and converted to a fraction of the total particles generated in each size bin using 10y/100 (Table 7.1). The fraction was applied to the total generation of particles (#/flush) and sampling volume (m3) averaged over the flush conditions (Table 7.1). Each size bin was corrected for its corresponding alveolar deposition efficiency for a breathing rate of 15 L air/ min, an 8 second breathing cycle, and 1 L of tidal volume as per (Heyder et al. 1986) . Ranges were specified using the nasal and oral deposition rates as the upper and lower bound, respectively. Exposure parameters for all toilet models are summarized in Table 7.2.
푛 퐷표푠푒퐿푒푔,푝푓 = ∑푖=1 퐶퐿푒푔퐶푎푒푟,푖 푉푎푒푟,푖퐼푡퐷퐸푖 Equation 7.3
Where DoseLeg,pf= the dose of Legionella deposited in the lungs per toilet flush; CLeg= the concentration of Legionella in bulk water; Caer,i =Generation rate/sampling
173 volume × fraction in size bin = the average concentration of aerosols of each diameter measured by Johnson et al. (2013); Vaer,i = the volume of each size bin of aerosols
3 calculated as V=(4/3)πr where diameter ranges from 1 to 10 µm; DEi= the alveolar deposition efficiencies of aerosols in each size range bin i; I= the mean inhalation rate of air breathed after toilet flushing (m3 air/ min); and t is the exposure duration or time spent in the room after toilet flushing (min).
174
Table 7.1 Aerosol size distribution for modern flush toilets (Johnson et al. 2013)
Toilet Type PAT, high-volume Pre-FEPA gravity HET, low Flushometer PAT, low-volume HET, high volume- flush flow volume- flush flush flush Liters per flush 4.9 13.3 3.8 5.3 4.2 4.9 Air sampling volume (m3) 0.013 0.01 0.012 0.0108 0.012 0.012 Total droplets produced (SE) 40,521 (1955) 54,363 (6764) 8220 (616) 145,214 (8325) 25,762 (1855) 10,620 (1060) Aerosols/ m3= 3.12×106 5.44×106 6.85×105 1.34×107 1.98×106 8.85×105 Total droplets/ Air Vol. Median droplet diameter (µm) Fraction* 1 1.49×10-2 1.37×10-2 1.27×10-2 1.21×10-2 1.31×10-2 1.14×10-2 2 1.29×10-2 1.24×10-2 1.18×10-2 1.09×10-2 1.24×10-2 1.11×10-2 3 1.11×10-2 1.13×10-2 1.07×10-2 1.04×10-2 1.04×10-2 1.07×10-2 4 1.07×10-2 1.11×10-2 9.90×10-3 1.02×10-2 9.89×10-3 1.07×10-2 5 1.06×10-2 1.13×10-2 1.03×10-2 1.04×10-2 1.00×10-2 1.08×10-2 6 1.06×10-2 1.11×10-2 1.08×10-2 1.02×10-2 1.01×10-2 1.09×10-2 7 1.06×10-2 1.09×10-2 1.07×10-2 1.02×10-2 1.02×10-2 1.08×10-2 8 1.05×10-2 1.08×10-2 1.05×10-2 1.04×10-2 1.01×10-2 1.05×10-2 9 1.04×10-2 1.07×10-2 1.04×10-2 1.04×10-2 1.02×10-2 1.04×10-2 10 1.02×10-2 1.05×10-2 1.04×10-2 1.03×10-2 1.01×10-2 1.05×10-2 *Fraction of total generated aerosols measured in each size bin Data were available for 1-10 µm size bins and were digitally extracted and averaged from each toilet type from Johnson et al. (2013) Figure 6, and converted to a fraction of the total particles generated in each size bin using 10y/100 175
Table 7.2 Exposure parameters for toilet flushing scenarios
Parameter Symbol Unit Value Distribution Source Model 1 Partitioning coefficient (PC) PC CFU m-3/ CFU L-1 1.3×10-6- 1.37×10-8 Uniform Barker and Jones (2005); Hines et al. (2014b) Fraction of total aerosolized F1-5 Unitless 0.75 Point Assumption in Schoen and organisms in aerosols of size range Ashbolt (2011) 1-5 µm Fraction of total aerosolized F5-6 Unitless 0.09 Point organisms in aerosols of size range 5-6 µm Fraction of total aerosolized F6-10 Unitless 0.14 Point organisms in aerosols of size range 6-10 µm Deposition efficiency of aerosols of DE1-5 Unitless 0.2 - 0.54 Uniform Schlesinger (1985) size range 1-5 µm deposited at the alveoli Deposition efficiency of aerosols of DE5-6 Unitless 0.1 - 0.65 Uniform size range 5-6 µm deposited at the alveoli
Deposition efficiency of aerosols of DE6-10 Unitless 0.01 -0.1 Uniform size range 6-10 µm deposited at the alveoli Model 2 3 a Concentration of aerosol in air after Caer,2.5 # aerosols/ cm air -1.246, 1.885 Lognormal O'Toole et al. (2009) toilet flush at 420 mm Deposition efficiency of aerosols in DE2.5 Unitless 0.41- 0.61 Uniform (nasal, oral) Heyder et al. (1986) alveolar region- 2.5 µm Model 3 b Concentration of aerosol : Caer,i Lognormal Johnson et al. (2013) 1 10.53, 0.87 2 10.43, 0.87 3 10.33, 0.89 4 10.30, 0.90 5 10.31, 0.90 6 10.31, 0.89 176
7 10.30, 0.90 8 10.30, 0.91 9 10.29, 0.91 10 10.28, 0.91 Deposition efficiency (Size) DEi Fraction 1 0.23, 0.25 Uniform (Nasal, Oral) Heyder et al. (1986) 2 0.4, 0.53 3 0.36, 0.62 4 0.29, 0.61 5 0.19, 0.52 6 0.1, 0.4 7 0.06, 0.29 8 0.03, 0.19 9 0.01, 0.12 10 0.01, 0.06 aLognormal distribution parameters shown are mean, standard deviation; bConcentrations of aerosol computed using average and standard deviation parameters across toilet types of Table 7.1 # Aerosols/ m3 * Fraction in size bin
177
7.7.6.2. Atmospheric dispersion model for cooling towers and spray irrigation
The primary types of atmospheric models for particle dispersion are simple box, Gaussian Plume (GP), Lagrangian, and Eulerian (Holmes and Morawska 2006).
Dungan (2010) and Van Leuken et al. (2015) review fate and transport models for bioaerosols, which have relied heavily upon modified GP models. QMRA models for wastewater, biosolids use, and spread of dusts containing pathogens between farms have used GP models with various modifications (Brooks et al. 2012, Brooks et al.
2005b, Dowd et al. 2000, Galada et al. 2012, Jahne et al. 2015, Jahne et al. 2014,
Ssematimba et al. 2012, Tanner et al. 2008, Teng et al. 2013, Viau et al. 2011), with several studies generating site specific, meteorological data-intensive estimates using
US Environmental Protection Agency AERMOD software (Dungan 2014, Jahne et al.
2015, Jahne et al. 2014) or CFD (Blatny et al. 2011, Blatny et al. 2008, Fossum et al.
2012). The goal of this work is to develop a generalized model for long-range transport with reclaimed water containing Legionella under a range of meteorological conditions that does not rely upon intensive site-specific information.
The concentration of Legionella downwind from an aerosol-emitting source is dependent upon the concentration of Legionella in the originating reclaimed water, transport and dispersion, deposition (wet and dry), evaporation, and bacterial viability as a function of environmental conditions. These factors are incorporated into a
Gaussian plume atmospheric transport model to calculate the dose of Legionella at a receptor downwind from a cooling tower or irrigation spray source using equation 7.4, a combination of previously proposed models that account for organism decay within the plume (Equation 7.4) (Lighthart and Mohr 1987, Peterson and Lighthart 1977,
Teltsch et al. 1980, USEPA 1982).
178
퐷표푠푒(푥, 푦, 푧) = 2 −휆 푥 푄 퐼푡 −푦 −(푧−퐻 )2 −(푧+퐻 )2 푠 퐿푒푔 푒 푒 푛 µ exp[( ) ] {푒푥푝 [ 2 ] + 푒푥푝 [ 2 ]} ∑푖=1 푞푖,푠퐷퐸푖푒푥푝 2훱µ휎푦휎푧 2휎푦 2휎푧 2휎푧
Equation 7.4
3 DLeg = Dose of Legionella at x, y, and z meters downwind from the source (# / m ), x
= distance downwind (m), y = horizontal distance perpendicular to wind (m) z = downwind receptor breathing zone height (1.5 m), He = Effective height of plume source from ground level (m) calculated as the maximum stream height for sprinklers
(Table 7.6) or the height of a cooling tower, µ = wind velocity [(m/s), determined by stability categories in Table 7.3], qi,s= the mass-weighted proportion of aerosols in each size bin 1 through 10 in the evaporated or aqueous aerosol state s (assumed to be
-1 uniform fractions); λ= microbial decay coefficient [s ]; σy = horizontal dispersion
3 coefficient (m), σz = vertical dispersion coefficient (m), I= the mean inhalation (m air/ min); and t is the exposure duration [min]. Downwind distances ranging from 50
– 10,000 m were simulated as this is the applicable range of the Gaussian plume model (Lighthart 1994). Exposure was assumed to occur in line with the plume centerline (y=0), which would be the maximum concentration distribution observed at a given x distance. Dispersion coefficients are calculated as per equations 7.5 and 7.6 where Ry, ry, Rz and rz are constants (Table 7.3).
푟푦 휎푦 = 푅푦푥 Equation 7.5
푟푧 휎푧 = 푅푧푥 Equation 7.6
QLeg is defined as per Equation 7.7:
푄퐿푒푔 = 퐶퐿푒푔퐹퐸 Equation 7.7
179
Where C= microorganism density in reclaimed water [organisms/L]; F= flow rate[L/s]; E= aerosolization efficiency= fraction of sprayed reclaimed water that leaves the immediate vicinity of the spray irrigation system as aerosols (0 180 Table 7.3 Pasquill Stability Classes for moderate solar radiation (Seinfeld 1986) Stability class Wind speed Ry ry Rz rz (Moderate (m/s) incoming solar radiation) A 1 0.469 0.903 0.017 1.380 B 3 0.306 0.885 0.072 1.021 C 5 0.230 0.855 0.076 0.879 D 7 0.219 0.764 0.140 0.727 The plume model accounts for dispersion, but not the fraction of aerosols within the respirable size range, which is of crucial concern for Legionella inhalation (1 – 10 µm). To obtain this fraction, the approach of Hauer (2010) was used, considering the mass-weighted fraction of aerosols that are likely to become fully evaporated as those that are < 100 µm in diameter, and the fine mist fraction as those droplets with diameters 100 – 200 µm in diameter (represented in the model as qi,s). Aerosols larger than 400 µm settled at distances < 50 m and were therefore assumed to settle at close range for both sprinklers and cooling towers and were not included in the model. It was assumed that all droplets with an initial diameter of < 200 µm would reach a diameter of 10 µm or less by the time they reached the downwind receptor. These fractions were reported by Hauer (2010) for various sprinklers. The largest fractions for Rainbird sprinkler nozzles were chosen (Table 7.5). The fine mist fraction was evenly divided into ten equal size bins to estimate the portion of downstream aerosols < 10 µm. Decay rates specific to evaporation or aqueous transport were applied separately for these portions of the total downstream aerosol load. For cooling towers, the mass-weighted proportions of aerosols < 100 and 100 - 200 µm were calculated by simulating a lognormal distribution specified by Peterson and Lighthart (1977) with a geometric mean and standard deviation diameter of 230 µm and 1.59, respectively. 181 Assumptions inherent in this model are 1) The background concentration of aerosolized Legionella spp. in ambient air is negligible; 2) Reclaimed water aerosols are generated during day time only (daytime solar insolation values and corresponding atmospheric stability values only are considered); 3) No overlapping irrigation sources or other sources of Legionella in the system; 4) Exposures occur at a constant distance directly downwind from the sprinkler or cooling tower; 5) Protection of Legionella due to the presence of organic debris or algae is not considered; 6) The impacts of aerosol dynamics including bubble burst, break up or agglomeration of aerosols, film collapse, and shear forces on Legionella are not considered; 7) effects of a moist aerosol plume thermodynamics are not considered; 8) no topographic effects; 9) no additional effects of biofilms and any biofilm with potential to slough off pipe surfaces was suspended in the bulk water at the time of sampling, therefore Legionella in bulk water represent 100% of Legionella available for aerosolization; 10) The fate of bacteria in individual aerosols is not tracked, however it is acknowledged here that larger aerosols in the starting distribution are likely to contain more bacteria and therefore result in higher concentration aerosols downstream than expected in some aerosols of smaller diameter (Blatny et al. 2011); 11) Enrichment of the aerosolized water with bacteria compared to the bulk water is not considered; 12) Reclaimed water is not blended with any other water source prior to use. Parameters used in the Monte Carlo analysis are summarized in Tables 7.4 and 7.5. 7.7.6.2.1. Spray irrigation Reclaimed water may be applied through a variety of mechanisms producing varying degrees of aerosols ranging from minimal (drip irrigation) to substantial (spray irrigation). Where biosolids are often distributed by moving tractors (Galada et 182 al. 2012), it is assumed here that spray irrigation with reclaimed water would take place via a stationary sprinkler system and could therefore be considered a point- source at 1 m in height or less. An example of common sprinkler systems used for reclaimed water production are provided in Table 7.6 (Jjemba et al. 2015). Sprinkler heights are all <1 m, however the plume model was considered to commence from the max height of the spray stream. The maximum stream height value of 6 m reported across the commonly used sprinklers in Table 7.6 was used. The distance required to reach the apex of the sprinkler stream (25 m) was added to estimates for sprinkler setback distances reported in the text. The sprinkler efficiency is the portion of initially sprayed water that leaves the immediate vicinity of the spray irrigation system as aerosols, including aqueous aerosols and evaporated droplets (USEPA 1982). Kohl (1974) measured the efficiency for low pressure smooth-plate sprinklers, which ranged from 0.5 to 1.4 %. 7.7.6.2.2. Cooling towers The principle categories for cooling water systems are once-through non- contact cooling, recirculation non-contact cooling, and direct contact cooling (Metcalf & Eddy 2007). The majority of cooling water systems that use reclaimed water are recirculating non-contact systems (Metcalf & Eddy 2007). In recirculating non- contact systems, warmed water, from a cooling operation or heat exchanger, is cooled by transferring its heat to air through evaporation in a cooling tower. Warm water from process cooling is sprayed on the top of the internal packing, used to break up the water through spray into droplets to enhance air/water contact. Cool, dry outside air is pulled up through the cooling tower by a large rotating fan to cool the warm water through evaporation. Water is removed by blowdown or purge, and small amounts of water capable of carrying microorganisms are also lost by drift. Typical 183 water loss from drift is assumed to be 0.001- 0.005% of the total recirculating water (ASHRAE 2004). However, it is noted that this loss rate could be higher for older designs or certain choices of drift eliminator (up to 0.1%) (Lucas et al. 2012). Both typical (0.003 to 0.005%) and high drift (0.01 to 0.1%) conditions were simulated to examine their impact on annual health risks. The design and operating conditions of recirculating cooling water towers vary widely (Selby et al. 1996). The flow rate of total recirculating water is a parameter designed (using performance curves specific to a given set of equipment and process being served) to achieve a desired range of thermal capability of the cooling tower, given set of operating conditions (entering water temperature, leaving water temperature, and entering air wet-bulb temperature) (ASHRAE 2004). The entering air wet-bulb temperature, required system temperature level, cooling tower size, and number of cells will balance the heat rejected at a specific approach (difference between leaving water temperature and entering air wet-bulb temperature) (ASHRAE 2004). The cooling tower size is a function of these factors as well as the quantity of water to be cooled, the air velocity through the cell, and the tower height (Zhang et al. 2012). Therefore, it is challenging to designate a set of typical operating conditions for use within a QMRA. It is assumed for simplicity in this model that larger heat loads necessitate larger towers, which in turn require larger quantities of recirculated water. Several studies report flow rates (per cell) for large (up to 100 m stack height) cooling towers of approximately 103- 104 L/s (Adams et al. 1978, 1980, Chen and Hanna 1978). Cooling towers of heights 10 m and 100 m were simulated with flow rates of 102-103, and 103-104, respectively. 184 7.7.7. Legionella concentrations in reclaimed water and drinking water Concentrations of Legionella in reclaimed water were log-transformed and a mean and standard deviation were computed for positive samples. The proportion of positive samples was incorporated as a separate variable into the Monte Carlo using a binomial distribution as some subsets of the data were highly censored (up to ~50% for culture measurements across all sites). 96% of cultured Legionella spp. and 52% of Legionella spp. gene copies were determined to be L. pneumophila. The recovery efficiency of the membrane filtration method was determined to be 70±18.6% (Mean ± SD). For all models, the concentration of Legionella in reclaimed water or drinking water CLeg was computed as per equation 7.8: 1 1 퐶퐿푒푔 = 퐶푅푊 푃푐표푛푡푎푚푓퐿푃 Equation 7.8 푅퐸 푛푠 Where CLeg= corrected concentration of Legionella pneumophila in reclaimed water or drinking water; RE= recovery efficiency; CRW= concentration of Legionella spp. measured in reclaimed water or drinking water (CDW); ns= number of samples for a given analytical method, per location (example: culture, all sites); Pcontam= binomial probability of contamination with number of samples ns and probability of contamination p; and fLP= fraction of Legionella spp. organisms in the analytical method identified as species Legionella pneumophila. A comparison of the current reclaimed water scenario with the use of potable water (“status-quo”) for toilets was made using concentrations of L. pnuemophila obtained from premise plumbing. Concentrations of Legionella spp. in the literature 185 from samples in homes or other buildings using culture-based methods ranged up to 1.25 x 106 CFU/ L (maximum of hot water samples) (Buse et al. 2012). Premise plumbing concentrations of Legionella were used for the toilet models where a drinking water scenario was conducted for comparison. A study of showers, taps, eyewash stations, fire sprinklers and recirculation loops conducted by Behets et al. (2007) provided these data. A mean and standard deviation of 2.85×103 ± 4.11×103 CFU/ L and 5.33×104 gene copies/ L was computed for positive samples using culture and qPCR-based methods, respectively. A comparison of the current reclaimed water scenario with the use of potable water for sprinklers and cooling towers was made using concentrations of L. pnuemophila obtained by Wang et al. (2012a). qPCR methods were used to measure L. pneumophila in a drinking water distribution system in Florida (Mean 90.4 GU/ mL ± SD 111.9, 5.6% water samples positive, n=54). Wang et al. (2012a) performed a regression analysis of qPCR and Legionella pneumophila culturing data in the study of Y=1.6X + 0.1 (R2= 0.9981), where Y= log (gene copies/ mL L. pneumophila) and X= log(CFU/ mL L. pneumophila). Legionella spp. was observed in only 1 of 56 samples (2%) by culture. If the regression equation is used to compute a point estimate culture value for L. pneumophila in potable water, the result is a mean of approximately 1.40×103 CFU/ L, which is comparable to concentrations detected by Morio et al.. Values from Morio et al. were therefore used for sprinkler and cooling tower comparison despite being from premise plumbing. The recovery efficiency of the study was not explicitly stated in any of the above studies, therefore the efficiency of the membrane filtration method observed in our study (70 ± 18 %) was used to correct all potable water concentration values. 186 Table 7.4 Concentrations distributions for Legionella in reclaimed and drinking water Parameter Symbol Unit Value Distribution Source Concentration of Legionella in CRW #/ L reclaimed water, positive samples only All locations and sampling points This study combined: Culture 10.003, 1.008 Lognormala qPCR- EMA 11.153, 2.169 qPCR- 23S 12.654, 2.008 By location: This study Site 4 Culture 10.074, 1.234 qPCR-EMA 11.426 , 2.887 qPCR-23S 12.616 , 2.347 Site 27 Culture 9.669, 0.623 qPCR-EMA 10.655, 2.691 qPCR-23S 11.621 , 2.936 Site 30 Culture 10.213, 0.829 qPCR-EMA 11.393 1.965 qPCR-23S 13.002 , 1.555 Site 31 Culture 9.898, 0.838 qPCR-EMA 11.584, 2.279 qPCR-23S 13.147, 1.756 Site 32 Culture 9.438 , 0.611 qPCR-EMA 11.139, 1.478 qPCR-23S 13.330 , 1.251 Site 33 Culture 10.183, 0.261 qPCR-EMA 10.318, 1.028 qPCR-23S 11.544, 0.912 (No sampling Q1) 187 Concentration of Legionella CDW, CT; Lognormal Wang et al. (2012a) pneumophila, municipal drinking CDW, sprinkler water distribution system qPCR gc / L 10.948, 0.964 Concentrations of Legionella pneumophila, premise plumbing Culture CDW, toilet CFU / L 7.393, 1.061 Behets et al. (2007); qPCR gc / L 10.228, 1.145 Masters and Edwards (2015) Culture CDW, CT CFU / L 6.574, 1.053 CDW, sprinkler b Portion of reclaimed water samples Pcontam (ns, p) Fraction Binomial This study contaminated with Legionella All locations and sampling points combined: Culture 153, 0.47 qPCR-EMA 115, 0.79 qPCR-23S 115, 0.80 By Location: Site 4 Culture 18, 0.78 qPCR-EMA 16, 1 qPCR 16, 1 Site 27 Culture 21, 0.52 qPCR-EMA 19, 0.74 qPCR 19, 0.79 Site 30 Culture 22, 0.36 qPCR-EMA 20, 0.55 qPCR 20, 0.55 Site 31 Culture 22, 0.77 qPCR-EMA 20, 1 qPCR 20, 1 Site 32 Culture 22, 0.59 qPCR-EMA 20, 1 qPCR 20,1 Site 33 188 Culture 22, 0.09 qPCR-EMA 20, 0.5 qPCR 20, 0.5 Portion of drinking water samples Binomial Wang et al. (2012a) contaminated with Legionella Culture Fraction 54, 0.056 qPCR Fraction 56, 0.02 Portion of Legionella observed that fLP % 96 Point This study is L. pneumophila for culture method Portion of Legionella observed that % 52 Point This study is L. pneumophila for qPCR method Recovery efficiency of membrane RE % µ= 70, 18.6 Normal This study filtration method (culture) a b Lognormal distribution parameters shown are mean, standard deviation; Binomial distribution parameters are number of samples (ns) and probability of contamination (p) 189 7.7.8. Decay rates Several studies have examined Legionella decay in aerosol as a function of relative humidity, seeding matrix, as well as bacteria strain, type, and source (Berendt 1980, 1981, Dennis and Lee 1988, Hambleton et al. 1983). Legionella survival generally increases as the ambient relative humidity increases, and it survives particularly well at intermediate relative humidities (65%). However, this relationship is not linear, and zones of instability are present. Legionella generally survives better in suspensions containing algal extracts compared to tryptose saline. Hambleton et al. (1983) held L. pneumophila 74/81 at various relative humidity for 15 minutes before aerosolizing them in a 3-jet collision nebulizer. The organisms survived best at 65% RH and worst at 55%. Survival was also high at 90% and 80% relative humidity. Therefore, 65%, 80%, and 90% relative humidities were chosen for modeling scenarios. Values were extracted from published graphs and decay constant K values were obtained by plotting log concentration versus time for two sets of experiments using water spray in Hambleton et al. (1983) (Table 7.5). Two other studies examined Legionella survival in aerosolized culture media broths (Berendt 1980, Dennis and Lee 1988) and were used as lower bounds on the decay estimates at each humidity. Only one study examined the decay of dried Legionella for use with the evaporated mass fraction (Katz and Hammel 1987). L. pneumophila Philadelphia 1 strain was dried for 90 minutes. A four-log drop in viability was observed during the first 30 seconds, followed by a more gradual decline. Biphasic K values were derived by converting percent recovery to a concentration and plotting log concentration versus time. The higher decay rate was applied for up to the first 30 seconds after the average 190 total evaporation time, while the lower decay rate was applied from t = 30 s to downstream time t where applicable. 191 Table 7.5 Monte Carlo model input parameters Parameter Symbol Unit Value Distribution Source 2 3 Flow rate of circulating water FCT, 10m L/ s 10 -10 Uniform Range of values for large towers 3 4 FCT, 100m 10 - 10 reported by Adams et al. (1978), Adams et al. (1980), Chen and Hanna (1978) 1 FIR L/s 0.329 Point Kincaid et al. (1996) Aerosolization efficiency ECT % Normal operating Uniform ASHRAE (2004) conditions: 0.001-0.005 Lucas et al. (2012) Less effective conditions: 0.1-0.01 EIR % 0.5-1.4 Uniform Kohl (1974) Horizontal distance y M 0; directly downstream Point Assumption perpendicular to plume along centerline of plume Downwind receptor breathing z M 1.5 Point Paez-Rubio et al. (2007) zone height Height of cooling tower HCT M Simulated for 10, 100 Point Assumption Height of irrigation sprinkler HIR M 6 Point assumed based Assumption on sprinkler characteristics (Table 7.6) Relative humidity RH % Simulated for 65, 80, and 90 Point Assumption -1 Decay of Legionella in aqueous λ1 s aerosol -5 -4 RH=65% λ1,65 8.40×10 - 2.38×10 Uniform Hambleton et al. (1983) -4 -4 RH= 80% λ1,80 1.82×10 - 3.09×10 Uniform Berendt (1980), Hambleton et al. (1983) -5 -4 RH=90% λ1,90 7.88×10 - 4.09×10 Uniform Hambleton et al. (1983), Dennis and Lee (1988) -1 Decay of evaporated Legionella λ2,t1 s 0.125 Point Katz and Hammel (1987) -4 (t1=up to 30s, t2= t- 30s if t > λ2,t2 3.10×10 Point 30s) Deposition efficiency DEi Fraction Uniform See Table 7.3, Model 3 Fraction of aerosols in respirable q (100 µm, Fraction range 200 µm 0.0138, 0.0413 (Rainbird 30 Point Hauer (2010) Sprinkler fractions) 5/32) 0.0459, 6.03×10-4 Point Peterson and Lighthart (1977) Cooling tower 1. Mean and lognormal parameters shown are mean, standard deviation. 192 7.7.9. Dose response and risk characterization The daily probability of each endpoint was calculated using the exponential dose response model for Legionella pneumophila (Equation 7.9) (Armstrong and Haas 2007a, Haas et al. 1999). −푟푑 푃푖푛푓 = 1 − 푒 Equation 7.9 Where Pinf is the probability of infection or clinical illness per event, r is the probability of the bacteria bypassing the host defenses and initiating infection, and d is the dose of Legionella at the target organ (alveoli). Annual risk was calculated as per Equation 7.10. 365푓 푃푖푛푓,푎푛푛푢푎푙 = 1 − ∏1 (1 − 푃푖푛푓) Equation 7.10 Where f is the daily frequency of the activity (flushing a toilet, spray irrigation application, or being present outside near a cooling tower). A sensitivity analysis was conducted to identify variables contributing to uncertainty using 104 Monte Carlo iterations. Risk characterization parameters are summarized in Table 7.7. The Spearman rank correlation coefficient was used to identify the most important predictive factors of annual infection or clinical severity infection risk, were 0 is no influence and -1 or +1 when the output is wholly dependent on that input. The model inputs were ranked based on their correlation coefficient with the output variable, annual risk. 193 Table 7.6 Risk characterization parameters Parameter Symbol Unit Value Distribution Source All models Inhalation rate, light activity, I m3/ min 0.013-0.017 Uniform USEPA (2011a) breathing cycle period 8 s and 1 L tidal volume Dose response parameter for L. r Unitless -2.934, 0.488 Lognormala Armstrong and Haas (2007a) pneumophila, infection endpoint Dose response parameter for L. r Unitless -9.688, 0.296 Lognormal Armstrong and Haas (2007a) pneumophila, clinical severity (death) endpoint Toilet flushing Exposure frequency f Flushes/ day 5.05, 2.69 Lognormal Mayer and DeOreo (1999) Exposure duration t Min/ flush 1-5 Uniform Lim et al. (2015) Spray irrigation Exposure frequency f Days / year Residential: 6 Point Brooks et al. (2005b), Brooks Occupational: 255 et al. (2012) Exposure duration t Hours / day Residential: 1 Point Assumption Occupational: 8 Cooling towers Exposure frequency f Days / year CT operates 255 days per Point Assumption year Exposure duration t Hours / day Residential: 1 Point Assumption Occupational: 8 aLognormal distribution parameters shown are mean, standard deviation 194 7.8. Results and discussion 7.8.6. Toilet flushing Three exposure model methodologies were used to compute the annual risk of infection and severe clinical infection from exposure to reclaimed water during toilet flushing. Annual risks for method 1 were highest, followed by method 3 and method 2. Annual risks for infection (infection endpoint dose response model, “Inf”) and clinical severity infection (death endpoint dose response model, “D”) overall and by site are shown in Figure 7.1. Annual risks did not vary substantially across locations. For data pooled across all sites (“All”) and median annual risks of infection (infection endpoint dose response model) and clinical severity infection (animal death endpoint dose response model) using culture-based Legionella concentration data ranged from 6.87 × 10-5 (method 2) to 9.32× 10-3 (method 1) and 7.83× 10-3 (method 2) to 1.10× 10-5 (method 1), respectively (Figure 7.3 and Table 7.8). Using PMA Legionella concentration data, median annual risks for infection and clinical severity infection ranged from 2.00× 10-4 (method 2) to 2.64× 10-2 (method 1) and 2.27× 10-7 (method 2) to 2.87× 10-5 (method 1), respectively. For qPCR Legionella concentration data, median annual risks for infection and clinical severity infection ranged from 8.34× 10- 4 (method 2) to 1.09× 10-1 (method 1) and 9.95× 10-7 (method 2) to 1.31× 10-4 (method 1), respectively. 95th percentiles for annual infection risks ranged from 3.20× 10-3 (method 2) to 1.08× 10-1 (method 1) for culture, 3.03× 10-2 (method 2) to 8.34× 10-1 (method 1) for PMA, and 1.14× 10-1 (method 2) to 9.99× 10-1 (method 1) for qPCR. 95th percentiles for annual clinical severity infection risks ranged from 3.09× 10-6 (method 2) to 1.20× 10-4 (method 1) for culture, 6.87× 10-5 (method 3) to 2.13× 195 10-3 (method 1) for PMA, and 1.42× 10-4 (method 2) to 7.52× 10-3 (method 1) for qPCR. 196 197 Figure 7.1 Annual risks for all and site-specific toilet flushing scenarios for infection (Inf) or Death (D) animal dose response model endpoints using three different exposure models. 198 If compared to the USEPA annual infection benchmark of 10-4 infections per person per year for drinking water, median annual infection risks (infection dose response endpoint) exceeded this value for all models except for method 2 / culture. Using a clinical severity infection dose response endpoint, only the median value from method 1, qPCR exceeded this value. 95th percentile annual clinical severity risks exceeded 1× 10-4 for method 1 all models, method 2 PMA and qPCR, and method 3 qPCR. Reclaimed water toilet flushing scenarios were compared to drinking water (status quo) toilet flushing scenarios in Table 7.8. In most cases, the use of drinking water containing Legionella pneumophila concentrations reported in the literature would not exceed a 1× 10-4 annual risk of infection (infection dose response endpoint). However, the 95th percentile for method 1 culture and qPCR, and method 2 qPCR annual infection risks exceeded this value. 95th percentile annual risks for drinking water did not exceed 1× 10-4 for clinical severity infection in any case. For culture data, median annual risks of infection or clinical severity infection for drinking water were 3 orders of magnitude lower than those for reclaimed water. For qPCR, median annual risks for drinking water were 1-2 orders of magnitude lower than for reclaimed water. 199 Table 7.7 Comparison of annual risks across all sites for toilet flushing scenarios using reclaimed water (R) or treated drinking water (DW) for infection (Inf) or Death (D) animal dose response model endpoints using three different exposure models Water Typea Infection Death Mean Median 5 95 Mean Median 5 95 Culture Method 1 R 2.60×10-2 9.32×10-3 5.46×10-4 1.08×10-1 3.13×10-5 1.10×10-5 6.95×10-7 1.20×10-4 DW 4.79×10-5 6.77×10-6 - 1.97×10-4 4.66×10-8 8.21×10-9 - 2.09×10-7 Method2 R 1.05×10-3 6.87×10-5 1.44×10-6 3.20×10-3 1.07×10-6 7.83×10-8 1.89×10-9 3.09×10-6 DW 1.23×10-6 3.30×10-8 - 3.83×10-6 1.67×10-9 4.20×10-11 - 5.57×10-9 Method3 R 1.10×10-3 4.05×10-4 4.10×10-5 4.10×10-3 1.23×10-6 2.76×10-7 1.70×10-8 4.86×10-6 DW 1.69×10-6 3.39×10-7 - 6.94×10-6 1.92×10-9 1.89×10-10 - 7.86×10-9 PMA Method1 R 1.44×10-1 2.64×10-2 3.15×10-4 8.34×10-1 7.37×10-4 2.87×10-5 4.13×10-7 2.13×10-3 Method2 R 1.01×10-2 2.00×10-4 1.25×10-6 3.03×10-2 3.89×10-5 2.27×10-7 7.67×10-9 1.42×10-4 Method3 R 1.66×10-2 1.16×10-3 1.95×10-5 6.84×10-2 3.38×10-5 7.83×10-7 9.26×10-9 6.87×10-5 qPCR Method1 R 2.81×10-1 1.09×10-1 1.66×10-3 9.99×10-1 2.24×10-3 1.31×10-4 2.08×10-6 7.52×10-3 DW 4.19×10-3 1.19×10-3 1.78×10-5 1.72×10-2 4.51×10-6 1.46×10-6 2.00×10-8 1.77×10-5 Method2 R 2.52×10-2 8.34×10-4 5.91×10-6 1.14×10-1 8.12×10-5 9.95×10-7 7.67×10-9 1.42×10-4 DW 1.28×10-4 8.28×10-6 3.01×10-8 4.35×10-4 1.61×10-7 9.76×10-9 2.11×10-11 4.77×10-7 Method3 R 4.52×10-2 5.30×10-3 1.03×10-4 2.30×10-1 8.66×10-5 3.40×10-6 5.41×10-8 2.19×10-4 DW 5.16×10-3 2.07×10-3 2.32×10-4 1.85×10-2 1.69×10-7 3.68×10-8 5.24×10-10 6.78×10-7 aR= reclaimed water, DW= treated drinking water 200 For all methods, the raw (uncorrected for recovery efficiency, portion of total organisms that are Legionella pneumophila, and fraction of samples contaminated) concentration of Legionella in reclaimed water is the most important predictive factor of the final estimate of annual risk of either infection or clinical severity infection (Figure 7.2). The large ranges of annual risks observed are therefore likely due to variability in this factor. For method 2, there was an exception for culture-based data, where the concentration of aerosols (Caer,2.5) was the most important factor. In all other cases for method 2, the concentration of aerosols was the next top-ranked factor. For method 1, the partitioning coefficient was the next most important factor. For all models, exposure time (t), dose response variable (r), and exposure frequency (f) were also important factors. 201 Figure 7.2 Sensitivity analysis showing Spearman rank correlation coefficients for toilet exposure model methods 1, 2, and 3. Coefficients identify the most important predictive factors of annual infection or clinical severity infection risk, where 0 is no influence and -1 or +1 when the output is wholly dependent on that input. 202 7.8.7. Cooling towers Annual health risks from exposure to aerosols from cooling towers were modeled for 4 Pasquil stability classes (A through D, corresponding to wind speeds ranging from 1 to 7 m/s), 3 humidity values (65, 80, and 90%), 2 stack heights (10 m and 100 m), 2 dose response endpoints (infection, clinical severity infection), 3 methods (culture, qPCR-PMA and qPCR 23S), and 2 exposure durations (residential, occupational) at various downwind distances from 50 - 10,000 m. A comparison of annual infection risks for residential populations with various combinations of wind speed and relative humidity parameters for Legionella culture data is shown in Figure 7.3. As wind speed increases, aerosols are carried farther, and annual risks peak farther away from the cooling tower. In addition, both types of annual infection risks are higher at a of stack height of 10 m compared to 100 m. Although Legionella is more stable at 65% and 90% relative humidity than at 80%, this does not have as great an impact on annual risk of infection as changing the wind speed. Legionella is carried farthest at the highest wind speed (7 m/s) and a relative humidity of 65%. Using this (most conservative) set of meteorological parameters, a comparison of annual risks for infection and clinical severity infection at two stack heights and all three analytical methods are shown for residential and occupational populations in Figures 7.4 and 7.5, respectively. Annual risks from qPCR are highest, followed by PMA and culture. Ninety-five percent confidence intervals for culture are more narrow than those for PMA and qPCR, likely due to the smaller variance of the Legionella concentration distributions for culture. Annual clinical severity infection risks are up to 2.5 orders of magnitude lower than annual infection risks at a given downwind distance. Occupational risks were up to one log higher than residential risks. Annual risks for cooling towers using drinking water were 2-3 orders of 203 magnitude lower than for using reclaimed water (Appendix Figure 9.10- Figure 9.11). Risks did not differ significantly by sampling location (Appendix Figure 9.12). 204 Figure 7.3 Annual risk (10y) of infection for Legionella pneumophila (culture method) in residential populations due to cooling tower exposure (stack height= 10 m, efficiency 0.001- 0.005%) at varying downwind distances and meteorological parameters. The median (solid black line) and 5th, 95th percentiles (dotted lines) are shown. 205 If the USEPA 10-4 annual infection risk target for drinking water is used for comparison to 95th percentiles from the cooling tower annual infection risk (infection dose response endpoint) distribution, the setback distance for both residential and occupational populations with a 10 m or 100 m tall cooling tower would be > 10,000 m with no other risk mitigation actions (Figure 7.4). Using a clinical-severity infection dose response endpoint and stack height of 10 m, setback distances for residential populations would cover a large range depending on the concentration method used. The setback distance would be less than 50 m for culture, ~2,000 m for PMA, or ~5,000 m for qPCR. For occupational populations, these corresponding distances would be ~1000 m, ~7,500m and > 10,000 m, respectively (Figure 7.5). If median residential annual clinical severity risks for a stack height of 10 m are compared to the 1×10-4 benchmark, only the qPCR values exceed this value and indicate a setback distance of ~500 m. For occupational exposure, median annual clinical severity risks for a stack height of 10 m indicate a setback of <50 m for culture, ~500 m for PMA, and ~2000 m for qPCR. For drinking water residential infection risks, setback distances of ~800 m for culture and ~5000 m for qPCR using a 95th percentile or < 50m and ~1000 m, using a median value, respectively, would be appropriate. Ninety-fifth percentile annual clinical severity risks for drinking water were <1×10-4 in all cases for a stack height of 10 m. 206 Stac Endpoint Method k Ht a (m) Culture PMA qPCR 10 Inf D 100 Inf D Distance from Source (m) aInfection (Inf) or Death (D) animal dose response model endpoints; death corresponds to clinical severity infection. Figure 7.4 Annual infection risks (10y) for Legionella pneumophila in residential populations due to cooling tower aerosol exposure (efficiency = 0.001-0.005%) at varying downwind distances for wind speed = 7 m / s and relative humidity = 65%. The median (solid black line) and 95% confidence interval (dotted lines) are shown 207 Stac Endpoint Method k Ht a (m) Culture PMA qPCR 10 Inf D 100 Inf D Distance from Source (m) Figure 7.5 Annual infection risks (10y) for Legionella pneumophila in occupational populations due to cooling tower aerosol exposure at varying downwind distances for wind speed = 7 m / s, relative humidity = 65%, and stack height=10 m. The median (solid black line) and 95% confidence interval (dotted lines) are shown 208 For a stack height of 100 m, culture and PMA annual clinical severity residential infection risks are <10-4 for all downwind distances. However, the 95th percentile annual clinical severity infection risk marginally exceeds 10-4 using the qPCR method for residential exposure. The 95th percentile value also exceeds the 10-4 target risk by 1-2 logs for both the PMA and qPCR method for occupational exposure. The 95th percentile annual clinical severity risks for drinking water were <10-4 in all cases for a stack height of 100 m and infection risks marginally exceeded 10-4 for qPCR for residential exposure and qPCR and culture for occupational exposure. The results of the sensitivity analysis for cooling towers are shown in Figure 7.6. Regardless of the analytical method used, the concentration of Legionella in reclaimed water was the most important predictive factor of the final estimate of annual infection or clinical severity infection risk. The cooling tower circulating water flow rate, dose response parameter r, and cooling tower drift efficiency were the next most influential factors. Due to variation in the efficiency of various drift eliminators, a comparison of “typical” operating conditions of 0.001-0.005 % and “less effective drift eliminator” conditions which might be typical of an older (pre-1970’s) cooling tower of 0.01-0.1% are presented in Appendix Figure 7.13. A higher efficiency drift eliminator can decrease risks by 1- 1.5 logs, highlighting this factor an important potential management strategy. 209 Figure 7.6 Sensitivity analysis for cooling towers 210 7.8.8. Spray irrigation Simulations for spray irrigation were conducted under the same conditions as for cooling towers and are presented for residential exposure under various meteorological conditions in Figure 7.7. Annual risks for drinking water for the same scenarios are shown in Appendix Figure 9.14 – 9.15. Similarly to cooling towers, the annual risks for sprinklers are affected by wind speed more than humidity, with a higher wind speed associated with longer transport of aerosols. A comparison of annual infection and clinical severity infection risks are shown in Figures 7.8 and 7.9. Annual risks did not differ significantly by sampling location (Appendix Figure 9.16). For sprinklers, a stack height of 6 m was used, corresponding to the highest point of the sprinkler spray arc. For this reason, Figures 7.7-7.9 and Appendix Figure 9.14- 9.16 should be interpreted as distance downwind from the horizontal distance at which the maximum height of the arc occurs. This distance has a maximum length of 24.4 m (Table 7.6) and is included in setback distance estimates below. In all cases, sprinkler risks were several orders of magnitude lower than cooling tower risks, most likely due to the smaller fraction of fine aerosols generated, lower dispersion height, and lower water flow rate. 211 Figure 7.7 Annual risk (10y) of infection for Legionella pneumophila (culture method) in residential populations due to sprinkler aerosol exposure at varying downwind distances and meteorological parameters. The median (solid black line) and 5th, 95th percentiles (dotted lines) are shown 212 Endpointa Method Culture PMA qPCR Inf D aInfection (Inf) or Death (D) animal dose response model endpoints; death corresponds to clinical severity infection. Figure 7.8 Annual infection risks (10y) for Legionella pneumophila in residential populations due to sprinkler exposure at varying downwind distances for wind speed = 7 m / s and relative humidity = 65%. The median (solid black line) and 95% confidence interval (dotted lines) are shown. 213 Endpointa Method Culture PMA qPCR Inf D Figure 7.9 Annual infection risks (10y) for Legionella pneumophila in occupational populations due to sprinkler aerosol exposure at varying downwind distances for wind speed = 7 m / s, and relative humidity = 65%. The median (solid black line) and 5th, 95th percentiles (dotted lines) are shown. 214 Table 7.8 Common sprinkler systems used for reclaimed water irrigation (NA=information not available from manufacturer) Irrigation sprayer Type Usage Device height Recommended Flow rate Spray radius Max stream Distance to (m) pressure range (L/s) (m) height (m) max spray (kPa) height (m) Rainbird Eagle Closed-case rotor Golf course fairways 0.09 410-690 1.35-3.60 19.2-29.6 6.1 18.3-24.4 900 Rainbird Eagle Gear-driven rotor Golf course roughs 0.07- 0.31 410-690 1.03-2.76 10.7-22.9 5.2 8.2-19.8 700 Toro 800 series Rotor Golf course 0.15-0.432 200-350 0.03-0.63 9.7-15.2 NA NA Hunter Pro-Spray Rotor (spray or Residential areas 0.05-0.3 100-700 0.01-0.36 2.6-5.8 NA 2.2-4.5 (spray head) rotating) Hunter PGP Rotor Residential areas 0.10 (total 206-482 0.032-0.91 6.7-15.9 2.1-4.0* 6.7-12.2* Rotors 4’’ device height 0.19) *At optimum operating pressures of 50-60psi; spray radius shown is across all operating pressures (https://www.hunterindustries.com/sites/default/files/PIG_PGP_dom.pdf) 215 Annual clinical severity residential population risks were <10-4 in all cases. For annual infection residential population risks, setback distances would range from 525 m (culture) to 10,025 m (qPCR) from the sprinkler under conservative meteorological conditions (conditions that promote Legionella dispersion of 7 m / s windspeed and 65% RH) and using a 95th percentile for comparison, if no other risk mitigation strategies are applied. These distances would be between <75m (culture) and ~2025 m (qPCR) using median values for comparison. Occupational annual infection risks indicate a setback distance >10,000 m and occupational clinical severity risks correspond to distances ranging from ~500 m (culture) to ~5025 m (qPCR) using a 95th percentile comparison and ~225 m (culture) to ~525 m (qPCR) using a median comparison. Irrigation with drinking water would result in annual risks < 1E-04 at all distances for residential exposure regardless of concentration method or dose response model used. For occupational exposures, setback distances would range from ~1025 m (culture) to ~5025 m (qPCR) if a 95th percentile is used for comparison and ~225- ~1025 m if a median is used for comparison. The sprinkler sensitivity analysis shown in Figure 7.10 also highlights the concentration of Legionella in reclaimed water as the primary factor that influences annual infection estimates. Dose response parameter r and the aerosolization efficiency were also identified as important factors. 216 Figure 7.10 Sensitivity analysis for sprinklers 217 7.9. Conclusions A QMRA is presented here for Legionella infection risks during toilet flushing, spray irrigation, and cooling tower mist inhalation with reclaimed water. Three toilet flushing exposure models were compared using different analytical methods and dose response endpoints. A modified Gaussian Plume dispersion model was used to compare risks for different analytical methods, meteorological conditions, exposure types, dose response endpoints, and downwind distances for sprinkler and cooling tower risks. All scenarios were compared to a drinking water “status quo” scenario to examine the relative risks of using reclaimed water over drinking water. In addition, two operation conditions for informing risk management were explored. These options were varying the stack height and aerosolization efficiency for sprinklers and cooling towers. Annual infection and clinical severity infection risks for Legionella did not vary substantially by sampling location for any scenario. Moderate differences were observed across the three analytical methods used to detect Legionella for each scenario (culture, PMA, qPCR methods). This resulted in 1-2 orders of magnitude differences in median annual infection or clinical severity infection risks across models. However, risk estimates for PMA and qPCR have higher values and a higher variance than risk estimates using culture-based data. This difference matters when choosing between percentiles to compare to the target or benchmark value in order to derive a setback distance, for example. Large differences (up to 3 orders of magnitude) across scenarios were observed when using an infection vs. a clinical severity infection dose response model, with the infection model resulting in higher risks. Lesser differences were observed in the cooling tower and sprinkler models at any downwind distance due to residential vs. occupational exposure (1-2 order 218 difference) or raising the stack height (up to 2 orders of magnitude difference, with peak risks occurring further downwind). Changes in meteorological parameters varied between models based on downwind distances, with high (7 m / s) wind speed and 65% humidity resulting in the highest risk estimates. The use of drinking water compared to reclaimed water resulted in up to 3 orders of magnitude lower annual risk estimates for all scenarios and combinations of other parameters. These findings indicate that the differences between risk scenarios (versus the sensitivity analyses that identify important factors within a single scenario) are most largely influenced by the selection of a dose response model (infection or clinical severity infection) or water source (reclaimed versus drinking water), followed by the exposure model used (toilet flushing only), exposure duration (cooling tower and sprinkler only; residential or occupational), and analytical method used to quantify Legionella in water sources (culture, PMA, or PCR). Annual infection or clinical severity infection risks greater than a tentative 10-4 annual infection benchmark value for drinking water were observed for certain cases in all three exposure scenarios, depending on the scenario conditions. For toilet flushing, method 2 was the least conservative (producing lower risk estimates overall) and method 1 was the most conservative (resulting in higher risk estimates overall). This is likely due to the use of a partitioning coefficient in method 1 which measures the ratio of bacterial concentrations in air and water and does not take into account the influence of aerosol size distribution. This approach is therefore likely to overestimate the proportion of bacteria in the respirable range at the receptor distances, especially as the distance at which the partitioning coefficient was derived is typically not available. Toilet risks can be mitigated by recommending reclaimed water users put the lid down prior to flushing. 219 The cooling tower and sprinkler models indicate that Legionella-containing aerosols can be carried long distances in sufficient quantities to present health risks above 10-4 annual probability of infection or clinical severity infection, in some cases past 10 km. These findings are consistent with previous studies that have predicted long-range transport of Legionella and observed distances between outbreak cases and implicated cooling towers up to more than 12 km (Borgen et al. 2008, Nygård et al. 2008, Rouil et al. 2004, Walser et al. 2014). However, these large spreads were likely produced by hot weather and high humidity, or rare events such as thermal inversions (Chan and Iseman 2013, Fisman 2005). In many cases, outbreaks were also associated with inadequate maintenance of cooling tower systems such as lack of regular inspection, faulty dosing pumps, suboptimal disinfection, high pressure cleaning, intermittent operation modes, and restarting of cooling towers (Walser et al. 2014). Depending on the model conditions selected, the setback distances associated with a 10-4 annual risk could be quite large. The setback distance was highly sensitive to the Legionella analytical method used. As a result, additional risk mitigation strategies are likely to be warranted to decrease the setback distance needed. Interventions such as windbreaks using trees or walls around irrigation areas could also reduce risks. Information is not currently available regarding the degree to which microorganism removal is achieved given these interventions. Risks from cooling towers can be reduced by utilizing towers built with a lower stack height and more efficient drift eliminators (Lucas et al. 2012). The simulations performed here demonstrate that a less-effective drift system can increase risks 1- 1.5 logs. Although higher stack heights have slightly lower annual risks than lower stack heights throughout the zone of influence, higher stack heights result in farther transport of aerosols and therefore higher setback distances. A lower stack 220 height reduces the distance that aerosols can travel. An effective stack height for cooling towers would be higher than the actual stack height used for simulation here due to the effects of plume rise, which occurs because the plume is hotter than the surrounding air and rises buoyantly as it exits the stack with a vertical velocity (Thomson et al. 2013c). In the current simulations, plume rise was not considered. In order to calculate plume rise, it is necessary to obtain specific information regarding the stack height exit velocity, stack diameter, and temperature of exiting water vapor, which required more specific information about the cooling tower than it is feasible to simulate given the coarse nature of the cooling tower height and circulating flow rate used in the current model. In general, for the low pressure, low profile sprinklers modeled, risks were much lower than for cooling towers. Annual clinical severity infection risks were <10- 4 for all scenarios with setback distances < 50 m. If considering annual infection risks, setback distances increase substantially beyond 50 m. In addition to using low-profile and low pressure sprinkler irrigation systems like the ones specified in this report, nozzles with larger orifices will reduce the formation of fine mist. Subsurface or drip irrigation would minimize drift formation, but can incur higher initial investment costs (Thomson et al. 2013d). Although increasing the droplet size distribution for sprinklers mitigates Legionella risk, reliance on larger droplets may increase erosion risk for fragile soils due to the greater kinetic energies associated with larger droplets (Montero et al. 2003). The sprinkler simulation is initiated at the apex of the spray arc. As such, the velocity propelling droplets in the x direction is not considered, which may cause larger droplets to be propelled further distances. This factor could result in actual 221 dispersion distances being greater than those used in the QMRA and therefore larger required setback distances. Cooling towers generally must operate when process cooling is needed, but sprinkler application periods can be chosen for during periods when there is low wind velocity or direction away from areas where sensitive populations are located such as hospitals. In addition, operations can be scheduled during nighttime hours or off- hours when employees are not scheduled to be present within the irrigation zone. Although not addressed in the models, conditions of high-humidity and lower temperature will reduce evaporative loss of droplets and prevent some of the size decrease that results in drift droplets reaching a respirable range before they settle. Although decay for evaporated Legionella is greater than for Legionella in aqueous aerosol, this factor had minimal influence in the model. Note that Legionella decay rates would likely be higher under daytime ambient environmental conditions compared to experimental conditions used by Hambleton et al. (1983) and others. Furthermore, it is not known how viability or infectivity evolve during aerosol transport. The most important factor identified in the sensitivity analysis for nearly all models was the concentration of Legionella in the reclaimed water. This analysis did not account for dilution of wastewater with other waters. Typically, reclaimed water can have up to 20 % v/v dilution (Dungan 2014). Dilution with water of higher microbiological quality could provide significant water savings benefits while still reducing the need for additional risk management options. Another important parameter identified in the sensitivity analysis is the aerosolization efficiency or fraction that ultimately is in the respirable range by the time it reaches a downwind receptor. Determining the aerosol size distribution and 222 downwind proportion of Legionella-containing aerosols in the respirable range remains a substantial challenge for QMRA models. It is challenging to model the evolution of aerosol size distributions over time due to co-occurring and interrelated dynamic rate physical phenomena of settling, evaporation, condensation, coalescence, and secondary aerosol formation due to bubble burst and film collapse (Hinds 1999, Lighthart et al. 1991). The fraction of aerosols in the respirable range is therefore not likely to remain constant over time and the current model may underestimate the impact of varying meteorological parameters as a constant fraction of respirable droplets over downwind distances is assumed. This approach has been applied in other Legionella QMRA models that considered aerosol size distributions (Armstrong and Haas 2007b, Nygård et al. 2008), however, a model that considers these factors is therefore recommended for further development to more accurately determine Legionella risks from systems with the potential for large scale dispersion. 7.10. Acknowledgements This work is supported by WateReuse Foundation Grant WRF12-05. 223 8. Conclusions 8.1. Summary The analysis undertaken in the current thesis work aimed to quantify microbiological health risks associated with opportunistic pathogens in roof-harvested rainwater. Based on a large field study and laboratory analysis using culture-based and qPCR methods in Brisbane, Australia, most opportunistic pathogens were prevalent in RHRW. The exception was L. pneumophila, which was rarely present (5% samples). The seasonal study of opportunistic pathogens in rainwater indicated that pathogens occur intermittently in RHRW, and display unique characteristics for individual pathogens. In all cases, fecal indicator bacteria were not correlated with opportunistic pathogens, indicating that current guidelines for determining rainwater safety are likely to be insufficient for opportunistic pathogens. A correlation analysis of pathogens with meteorological factors indicated that while fecal indicator bacteria such as E. coli or Enterococci were related to rainfall, this was not the case for all of the opportunistic pathogens. A relationship between P. aeruginosa and stagnation time emerged, indicating that management strategies might differ for fecal-associated bacteria and opportunistic pathogens in RHRW tanks, as increased stagnation time is often recommended for attenuating fecal-associated pathogens but might lead to increased growth for opportunistic pathogens. A quantitative microbial risk assessment (QMRA) of opportunistic pathogens L. pneumophila and MAC in RHRW indicated that some uses may exceed common risk benchmarks. In reclaimed water, Legionella species and free-living amoeba were found frequently. While Legionella infection risks did not vary substantially by sampling location, the choice of microbiological method (culture, qPCR, or EMA-qPCR) for concentration input data had a large impact on risk estimates, and therefore setback distance determinations. Despite the small volume inhaled for toilet flushing scenarios, risks still exceeded annual risk benchmarks for some cases. Toilet lids can be closed in residential homes but in certain public locations, 224 toilet lids are not present and this is not a risk management option. In many cases, large setback distances would be necessary for sprinklers and cooling towers to achieve risk benchmarks; other management options such as wind-breaks, trees, or increased treatment of reclaimed water can decrease the distance necessary to protect public health. 8.2. Future work Several areas for further study have been identified in the current work. First, the field studies of opportunistic pathogens in RHRW performed in this thesis work were conducted using qPCR assays. These assays provide information on the number of gene copies per liter of water, but do not provide information regarding viable and/or infective cells. This information is important for risk assessments as demonstrated in chapter 7 is likely responsible for the large difference in risks computed between qPCR and culture methods. However, culture-based methods will not quantify microorganisms in a VBNC state and therefore are limited. Additional study is warranted regarding the proportion of viable/infectious pathogens, especially with regards to seasonal factors. Additionally, in the rainwater studies, correlations were performed between pathogen concentrations and meteorological factors. However, additional work could explore the impact of survey factors on pathogen concentrations. Other studies have demonstrated the importance of biofilms for harboring opportunistic pathogens; future rainwater work should focus on biofilms in these systems. Biofilm work was not pursued in this study due to the difficulty in obtaining a biofilm sample from the tanks; the inside of in-use tanks was not easily accessible, and cleaning companies that accessed the inside of tanks did not sterilize their equipment, introducing the potential for inoculation with foreign microorganisms if tank biofilms are examined during the cleaning process. For MAC dose response models, the fit by Tomioka et al. is arguably the most important model fit because of its endpoint of pulmonary disease. However, two points in this model are not encompassed by the 99% confidence interval of the bootstrapped model fit. 225 Therefore, future work can focus on the development of multi-hit or other, “steeper” models that will better encapsulate the data points. Additionally, the development of further models for immune-compromised hosts will be beneficial for future assessments of MAC risk. For pathogen QMRA, all risks were reported in terms of annual probability of a given endpoint. However, it is recommended that future work calculate these probabilities in terms of disability adjusted life years (DALYs) for comparison with World Health Organization tolerable risk levels. For long-range Legionella QMRA, the modified Gaussian plume model used in the current thesis work uses static proportions of aqueous aerosols in respirable size bins. However, aerosol sizes will evolve as droplets migrate further from the source. A dynamic aerosol size model could be compared to the current static model to determine the extent to which aerosol size distribution changes impact Legionella infection risks. 226 9. Appendix 9.1. RHRW literature review Table 9.1 Literature review search results for RHRW (Q=queried, I= imported to digital library) Proquest Agricol Engineering Village Google Web of Environmental a Scholar Knowledge Engineering Abstracts Q I Q I Q I Q I Q I "harvested" AND "rainwater" AND 6 5 3 3 16 16 21 12 14 14 "quality" "harvested" AND "rainwater" AND 0 0 6 3 17 10 2 2 3 3 "risk" "harvested" AND "rainwater" AND 2 2 8 7 13 12 3 1 0 0 "bacteria" "harvested" AND "rainwater" AND 2 2 0 0 1 1 0 0 2 2 "microorganism" "harvested" AND "rainwater" AND 0 0 0 0 0 0 0 0 25 22 "protozoa" "harvested" AND "rainwater" AND 0 0 0 0 1 1 0 0 3 3 "pathogen" "harvested" AND "rainwater" AND 0 0 1 1 5 4 1 1 6 5 "review" "harvested" AND "rainwater" AND 0 0 2 2 3 3 1 1 3 3 "qmra" OR quantitative microbial risk assessment "harvested" AND "rainwater" AND 8 7 6 6 91 44 4 2 72 28 "chemical” "harvested" AND "rainwater" AND 4 4 2 1 60 28 0 0 48 25 “metal” "rain" AND "water" AND "quality" 72 2 10 3 17 7 71 13 13 14 "rain" AND "water" AND "risk" 32 1 0 0 5 2 2 0 1 0 "rain" AND "water" AND "bacteria" 18 4 1 0 0 0 4 0 475 19 "rain" AND "water" AND 10 2 0 0 2 1 0 0 15 0 "microorganism" "rain" AND "water" AND 1 0 0 0 0 0 0 0 25 1 "protozoa" "rain" AND "water" AND 5 2 0 0 1 0 0 0 95 3 "pathogen" "rain" AND "water" AND "review" 12 1 3 0 3 0 8 0 95 5 "rain" AND "water" AND "qmra" 0 0 0 0 0 0 0 0 3 2 "rain" AND "water" AND 98 6 620 9 22 0 78 4 22 1 "chemical" "rain" AND "water" AND "metal" 42 4 76 3 4 0 11 0 19 0 "rainwater" AND "quality" 45 14 15 1 98 63 17 53 632 157 2 5 "rainwater" AND "risk" 9 2 0 0 10 7 16 7 240 34 "rainwater" AND "bacteria" 6 4 0 0 3 3 9 4 125 34 "rainwater" AND "microorganism" 7 4 0 0 2 1 0 0 3 0 "rainwater" AND "protozoa" 0 0 0 0 0 0 0 0 7 4 "rainwater" AND "review" 3 2 0 0 12 5 20 1 85 11 "rainwater" AND "qmra" 0 0 0 0 0 0 1 1 4 3 "rainwater" AND "pathogen" 2 0 0 0 4 4 1 0 22 4 “rainwater” AND “chemical” 64 6 114 5 88 5 12 3 88 9 6 227 “rainwater” AND “metal” 46 7 25 2 35 3 10 2 539 98 "cistern" AND "quality" 2 2 3 0 8 7 13 2 68 14 "cistern" AND "risk" 1 1 0 0 3 2 1 0 129 6 "cistern" AND "bacteria" 0 0 0 0 0 0 0 0 41 9 "cistern" AND "microorganism" 0 0 0 0 0 0 0 0 3 0 "cistern" AND "protozoa" 0 0 0 0 0 0 0 0 2 1 "cistern" AND "pathogen" 0 0 0 0 0 0 0 0 4 1 "cistern" AND "review" 0 0 0 0 1 1 8 0 143 2 "cistern" AND "qmra" 0 0 0 0 0 0 0 0 0 0 “cistern” AND “chemical” 2 1 0 0 45 4 0 0 95 7 “cistern” AND “metal” 0 0 1 0 37 9 0 0 30 4 "roof" AND "water" AND "risk" 1 0 0 0 6 2 7 2 80 20 "roof" AND "water" AND "bacteria" 4 3 0 0 0 0 1 0 42 18 "roof" AND "water" AND 5 4 0 0 0 0 0 0 0 0 "microorganism" "roof" AND "water" AND 0 0 0 0 0 0 0 0 6 4 "protozoa" "roof" AND "water" AND 1 1 0 0 0 0 0 0 5 4 "pathogen" "roof" AND "water" AND "review" 6 6 0 0 4 2 2 1 42 6 "roof" AND "water" AND qmra" 0 0 0 0 0 0 0 0 1 1 “roof” AND “water” AND 23 9 14 7 3 3 3 2 289 90 “chemical” “roof” AND “water” AND “metal” 26 12 6 2 4 3 5 1 210 109 "roof" AND "water" AND "quality" 29 15 0 0 34 20 55 24 287 98 228 Table 9.2 Summary of design and maintenance for reviewed rainwater collection apparatuses Reference Type* Setup First Flush Alsup et al. (2011) Edwardsville, P 3 green roof models used: unfilled (no root barrier), Built in Place (BIP; metal side walls, drainage No first flush diversion; covered gutter Illinois layer, root barrier system, and roofing membrane adhered to wafer board substrate), and block model collects leachate (commercial aluminum inside water retention box with drain to collection container; same depth as BIP, fertilized). Birks et al. (2004) F Rainwater collected from Millenium Dome at Thames Water Recycling Plant (100m3/day) No first flush diversion London, UK Clark et al. (2008) P, L Pilot roofs are panels with shingles attached in sheet rows attached to A frames with slopes similar to No first flush diverted, pilot scale NR Alabama and Pennsylvania, US residential roofs. Plexiglass used as a control for subtracting background concentrations. Domènech and Saurí (2011) F NR NR Suburb of Sant Cugat de Vallès, Barcelona, Spain Farreny et al. (2011) Barcelona, Spain F Metal sheet (smooth, 30° slope), polycarbonate plastic (smooth, 30° slope) , clay tile (slightly rougher, No first flush diversion 30° slope), and flat gravel roofs (rough, 1° slope, 15-20 mm gravel depth, ~5mm particle diameter). Roof footprints 40.6m2- 120m2. No maintenance conducted over study period. Gikas and Tsihrintzis (2012) F Clay tile (30°, 100-180m2 surface area, 1000-3000 L total storage tank volume), concrete (1°, 75- Clay 11-20 L first flush (0.11- 0.12 mm); Xanthi, Greece 100m2, 1000-2000 L), or maxitherm (30°, 180m2, 2000 L) in rural, suburban, urban or university concrete 10-13 L (0.13 mm); maxitherm campus areas. none diverted Kelly et al. (2011) Lake Ontario, F, D Copper or non-copper roofs No first flush diversion Canada Madonia et al. (2012) F Site of persistent volcanic activity. Limewashed roof surfaces with metal/plastic vertical collection Manual bypass removes first flush; Stromboli Island, Italy (volcanic pipe and underground sloped-floor cistern lined with limewashed bricks. Cistern residence time > 1 volume unspecified. archipelago) hydrological year. Mendez et al. (2010) and Mendez et P, D Pilot scale roofs: Ashphalt fiberglass shingle and concrete tile (18.4°, 2.8 m2), unfertilized green roofs Passive collection system with 2 L first al. (2011) (type E), and 2-ply atactic polypropylene modified bituminous membrane cool roofs (1.2°, 3.4m2). flush; all measurements post first-flush. Austin, TX Full-scale roofs: residential 12-year-old Galvalume® (22°, 4.3m2), 5-yr-old asphalt fiberglass shingle (23°, 4.3m2, and asphalt fiberglass shingle roof (18° 5.3m2). O'Hogain et al. (2012) P, F Rainwater from two sheds/barns (1000m2) gravity drained to underground 9m3 concrete tank. Steel No first flush diversion; Clonalvy, Ireland (livestock farm 50 mesh placed in downpipe gutter as a filter. Rainwater pumped to concrete reservoir tanks where km from Dublin) sampled. Poster and Baker (1996) D Automated wet-only sample collector with in situ filtration system . NA Maryland, US Schets et al. (2010) P, F, L Vegetated (untreated) , rubber (not vegetated, filtered), and unspecified roof/ balcony (filtered or NR The Netherlands unfiltered) roof -runoff sampled from underground reservoir outside buildings; in use since 1996-2001. Schriewer et al. (2008) F 14 year old zinc roof (10°, 238m2). Roof panels and gutters made of titanium-zinc. Five small chimneys Sampling process starts after 60s with Munich, Germany soldered on roof with tin-solder containing fractions of lead. flow ≥ 1.0 L/ min; stopped after another 60s of flow < 0.6 L/min. Stump et al. (2012) F Samples from storage tanks (fiberglass, polypropylene, steel, or ferrocement) or faucet between tank 21/36 (58%) systems have first-flush Texas, US and purification system (9,463-189,270 L; mean 75,708 L). Wood, steel, concrete tanks have food- diversion but some less than recommended grade liners. Roofs were metal (32/36), 3 with composite shingles, 1 with clay tile. 76L / 93m2 roof area. Teemusk and Mander (2007) F Green roof (2 outflows) with modified bituminous base roof, plastic wave drainage layer (8mm), rock No first flush diversion 229 Tartu, Estonia wool for rainwater retention (80mm), and a substrate layer (100mm) with LWA (66%), humus (30%) and clay (4%). Reference roof (1 outflow) is a modified bituminous membrane roof. Both roofs 0°, 120m2. Plant cover 45%. Vialle et al. (2011) and Vialle et al. F 204m2 tiled roof with open zinc gutters, mesh filter, and underground 5m3 storage tank (samples at No first flush diversion (2012) water surface). Submerged intake with inlet filter used to pump water inside house. Southwest France *P= Pilot Scale; B= Bench (laboratory) scale; F=field scale or actual installation in use; D= direct rainfall; NR= Not reported; NA= Not applicable 230 9.1.1. Factors that impact RHRW quality 9.1.1.1. Sources of contamination Microorganisms can enter RHRW containers through aerosol deposition, tree litter, and animal fecal matter (Brodie et al. 2007, Evans et al. 2006, Huston et al. 2009, TWDB 2005, Whon et al. 2012). Metals or chemicals can originate from airborne soil or industrial emissions, corrosion of roof materials and dissolution from sediments in storage chambers or the distribution system (Lye 1992). While falling, atmospheric contaminants can be absorbed including sulfates (SO4), ammonium + - 3- (NH4 ), nitrates (NO3 ), and phosphates (PO4 ) (Gobel et al. 2007). Rain also absorbs carbon dioxide and nitrogen (TWDB 2005), sulfur oxides (SOx), nitrogen oxides (NOx), and chloride (Cl) from combustion installations, causing rain to be acidic in some cases (Gobel et al. 2007). Volatile organic carbons (VOCs) from airborne chemicals associated with agricultural and industrial processes can also be absorbed (White 2007). 9.1.1.2. RHRW system design The materials associated with the construction of RHRW systems are particularly important for water quality. Heavy metals are present in RHRW through their release from roofing, gutter, or pipe and faucet materials. Galvanized iron roofs as well as wooden shingle roofs have been identified as prime sources of zinc, as zinc is found in the galvanizing material and zinc sulfate and zinc chloride mixtures are used to deter moss growth on wood single roofs (Chang 2004, Forster 1999, Macomber 2004). A comparison of wood shingle, composition shingle, painted aluminum and galvanized iron roofs found zinc concentrations to be highest in runoff from the wood shingles, followed by galvanized iron, painted aluminum and 231 composition shingles (Chang 2004). Copper and lead concentrations in runoff from roofs with copper flashings have been noted to be six to eight times greater than galvanized roofs (Forster 1999, Macomber 2004). Old copper pipes and brass faucets also often contain lead solder which provides an additional source of lead. Acidic rain further contributes to the release of heavy metals from these sources (Macomber 2004). A comparison of asphalt fiberglass shingle, metal, concrete tile, cool, and green roof materials by Mendez et al. (2011) showed that the first flush from metal roofs tended to have lower concentrations of fecal indicator bacteria. Although shingle and green roofs produced water of comparable quality, high dissolved organic carbon concentrations could result in disinfection by-products. Furthermore, high concentrations of metals present in runoff from green roofs highlights the importance of soil-media selection. Similarly, Lee et al. (2012) found metal roofs to be the most suitable for RHRW based on total suspended solids, nutrients, metals, and E. coli concentrations. A review by Rowe (2011) demonstrates some conflicting reports on the effectiveness of green roofs for pollutant removal but highlights roof age, choice of vegetation, fertilization and maintenance, local pollution sources, seasonality and local rainfall as important factors. The use of alternative roofing materials or roof coatings materials to lower pollutant loads has also been suggested and research in this area regarding TiO2 coatings have been pursued (Kim et al. 2004). Various filtration materials and absorber systems have also been developed and are aimed at the removal of copper and zinc (Athanasiadis et al. 2007, Athanasiadis et al. 2004, Macomber 2004, Steiner and Boiler 2006). The installation of detention basins or sediment traps have also been reported to potentially act as an effective method for the elimination of nutrients and 232 oxygen demanding materials (Jackson 2003). Across the 16 reviewed studies, half did not remove the first flush before measuring water quality (8/16) or measured direct rainfall water (3/16). First flushes removed (where volume was reported) ranged from 2 to to 20 L (3/16), however some studies did not specify the volume diverted (2/16) or did not specify whether the first flush was removed (2/16). 9.1.1.3. Environmental factors The composition of pollutants that are deposited on catchment surfaces is highly variable depending on the location of the roof catchment area. In urban areas of Germany, it was found that background air pollution and local sources (PAHs from heating sources) played a great role in determining the quality of harvested rainwater (Forster 1999). Physical and chemical rainwater quality is typically of lower quality in urban areas (Marciano-Cabral et al. 2010, Melidis et al. 2007). Precipitation events, hydrologic, and meteorological factors also have seasonal and diurnal patterns affecting harvested rainwater quality (Scholz 2004). Low rainfall intensities have shown to be correlated with higher Zinc concentrations (Schriewer et al. 2008). Rainfall intensity has also been shown to affect heavy metal speciation (Madigan et al. 2012). In general, decreasing concentrations of some chemicals (chloride, sulfate, sodium, calcium, nitrate) were observed as average cumulative rainfall depths increased (over the course of individual storm durations), likely due to atmospheric ‘washout’ and removal of dry materials on the catchment surface (Evans et al. 2006). In addition, the bacterial load in RHRW is directly affected by wind speed while the composition of microorganisms varies with wind direction (Evans et al. 2006). 233 During dry antecedent periods, dusts, aerosols and gas from the atmosphere can be directly transferred by deposition to tanks if openings exist (Gobel et al. 2007). The duration of dry antecedent conditions plays a role in the amount of accumulated pollutants that are washed from roofs into rainwater containers, with longer dry antecedent periods associated with higher bacterial counts (Forster 1996, 1999, Gobel et al. 2007, Leckie 2005, Schriewer et al. 2008, Yaziz et al. 1989). However, some parameters such as zinc concentrations have been found to be independent of antecedent conditions (Schriewer et al. 2008). 9.1.2. Impacts of RHRW use 9.1.2.1. Environmental impacts The use of RHRW is an important adaption to climate change fluctuations and provides opportunities for stormwater management (Pandey et al. 2003). However, roof runoff can also be a source of pollution. In a study of storm water pollution in Washington, DC, Sharifi et al. (2014) estimated that roof runoff can be responsible for up to 50% of lead in storm water. In addition, irrigating with water high in pathogens, chemicals, nutrients, and metals could create agriculture or plant toxicity. High levels of oxygen demanding materials trapped within the soil layer could lead to the depletion of the oxygen levels in soil water matrix. In extreme cases, this leads to the die-off of aerobic nitrifying soil microbes and toxicity to vegetation through the mobilization of metals in the soil (Sawyer 2003). High levels of total dissolved solids can indicate a potential for plant mortality or affect its development, as it inhibits the plant’s ability to uptake water by decreasing the osmotic gradient in the root zone (Asano 2007). 234 Soils with pH values between 5.5 and 8.0 will precipitate ions and eliminate toxicity (Fipps 2003). Rain absorbs carbon dioxide as it falls, manifesting in a form of carbonic acid which gives the rain slight acidity (roughly 5.5) when exposed to the atmosphere with little buffering (Sawyer 2003). Manganese is toxic to a number of plant species at a few tenths to a few mg/L, particularly in acidic soils. Vanadium can also be toxic to many plants even at relatively low concentrations (Asano 2007). Metals generally exist in soils in insoluble forms that are immobile and not readily taken up by plants, lending to their relatively low threat of toxicity. However, dissolved metal species can exist as free metal ions, inorganic complexes and organic complexes, with varying levels of bioavailability and toxicity (Brown et al. 2005, Stumm and Morgan 1996). Under acidic or highly anaerobic reducing conditions, metals can become mobilized, posing threats both to vegetative species and potentially groundwater (Sawyer 2003). Additionally, the uptake of metals and other constituents of concern in plant tissue over time and the subsequent health risks associated with consumption have not yet been fully characterized, presenting an area for further evaluation and research (Sayyad et al. 2009). 9.1.2.2. Human exposure Rainwater use is practiced in North America and Europe and presents known exposure routes for contaminants present in RHRW. However, limited information is available on actual end uses and their frequencies, and whether these differ from use patterns traditionally observed for municipal water sources alone (Mayer and DeOreo 1999). Some potential RHRW exposure scenarios include potable use, non-potable reuse (toilet flushing, household cleaning, clothes washing, lawn irrigation, 235 emergency supplies for fire-fighting, ornamental use), irrigation of produce for consumption, cooking, and occupational exposures via installation and maintenance processes (Hatt et al. 2004, Lye 1992, Struck 2011). In other developed and developing countries, rainwater harvesting and reuse have been widely practiced for potable and nonpotable applications such as irrigation and stormwater management purposes. Many regions such as Australia, East Africa, Zambia, China, Singapore, Greece, and Bermuda rely on harvested rainwater as a primary source of water for one or more of these uses (Lye 2002). RHRW has been found suitable for drinking water in some cases in these areas, although exceedances of health based guidelines for drinking water have been observed (Handia 2005, Jackson et al. 2001, Melidis et al. 2007, Peters et al. 2008, Sazakli et al. 2007, Thomas 1998). A few studies have provided information on RHRW usage in the United States. A 1991 survey of American State and Territorial Health Departments by Lye (1992) estimated that 214,525 cistern systems were in use in the US (and territories). The states with the greatest numbers of reported private cisterns were Kentucky (~80,000), Ohio (~67,000), Virginia (~35,000), Tennessee (~15,000), Alaska (~10,000), and Hawaii (~7,000) (Lye 1992). Maintenance of these systems was not well studied and could be prone to fecal contamination from birds or rodents (Lye 1992, USEPA 1984). Thomas et al. (2014) conducted a survey of individual and business members (n=136) of the American Rainwater Catchment Systems Association (ARCSA), representing 2700 RHRW systems across the United States that the respondents had owned, installed, or observed in the field. The most commonly used materials were asphalt shingles and metal for the roofing material, polyethylene for the cisterns, and seamless aluminum for guttering and downspouts. Individual potable water users 236 used primarily metal roofs (86%). The most common uses for harvested rainwater were irrigation (90%), non-potable indoor use (43%), and drinking water (32%). 21% of individual respondents and 37% of business respondents did not conduct microbial water quality testing. For chemical testing, 30% and 32% of individual and business respondents did not conduct testing, respectively. Studies at smaller regional scales indicate rainwater use for toilet flushing, garden irrigation, and vehicle washing in the US (North Carolina, Oregon, Texas), and Canada (Baird et al. 2013, Crowley 2005, Jones et al. 2006, Jones and Hunt 2010, Stump et al. 2012). Stump et al. (2012) surveyed 36 households who were members of the Texas Rainwater Catchment Association and had rainwater collection systems intended for indoor potable use. RHRW was the sole water source for 56% households. Thirty-two of these homes had metal roofs, and although 58% had a first flush diversion mechanism in place, some did not meet recommended guidelines for first flush:roof area (76 L: 93 m2 roof material) (Stump et al. 2012). The majority (64%) of these household drinking water systems had never been tested for contaminants. 237 Table 9.3 Rainwater harvesting practices from 136 survey participants in the United States adapted from Thomas et al. (2014) Parameter Category % Individual users who % Business users who reported (n=128 reported (n=136, 36 for respondents/ 2700 RHRW systems, 51 potable water uses) businesses/1047 systems for potable water; percentages corrected for number of systems per respondent) Location Located above ground 61 57 Buried cistern 35 Water drawn from spigot near 63 cistern bottom System age (yrs) 0-2 41 3-5 32 5-10 18 11+ 9 Motivation for RHRW Cost effective 34 25 Self sufficiency 8 13 Help environment 4 2 Use of RHRW Irrigation 90 94 Non-potable indoor 43 69 Drinking water 32 39 Roof material Asphalt shingle 48 (n=158) 29 (n=2343) Metal 38 46 Other 6 Clay/concrete tile 5 16 Cool 3 2 Green 2 Not sure 2 Cistern material Polyethylene 67 (n=159) 51 (n=2599) Concrete 10 12 Galvanized 10 14 Other 9 2 Fiberglass 3 11 Wood 1 Pioneer/Modular 6 Not sure 4 Gutter/downspout material Seamless aluminum 66 (n=158) 59 (n=2587) 238 Other 20 8 PVC 12 24 Not sure 2 5 Galvanized steel 4 Cistern size, gallons <1,000 22 (n=143) 4 (n=2423) 1,000-2,000 33 28 3,000-5,000 15 20 6,000-10,000 14 14 11,000-20,000 11 28 20,000+ 5 6 Treatment Divert first-flush 51 54 Roof washer (filter upstream of 67 73 cistern) Backup water supply Municipal water 63 52 Private groundwater 19 15 Water truck delivery 5 12 No backup 13 21 Drinking Water users (n=36 individual users, 1047 systems represented by business users) % Potable use by roof type Asphalt shingle + small cistern 6 <2000 gal (n=51) Metal (n=36) 86 Filtration method (multiple selections possible) Combination of treatment 70 (n=50) 79 (n=1418) approaches Not filtration alone 32 No filtration 6 Cartridge filter 26 48 GAC filter 20 39 Other 18 Reverse osmosis 4 7 Disinfection method In-Line UV treatment 75 (n=40) 70 (n=898) In-line chlorination 4 Other 18 15 Batch chlorine in cistern 5 4 Not applicable 2 Not sure 7 Storage Pressurized tank 28 24 Unpressurized tank 31 38 No storage, treated on-demand 41 38 Microbial testing frequency Not tested 21 37 (n=680) 239 Not sure 6 Yearly 46 50 Every 6 months 15 10 Every 3 months 12 3 Water quality testing type Not tested 22 24 Commercial company 28 38 Government agency 31 25 Home test kit 14 13 Other 5 Chemical testing frequency Not tested 30 32 Not sure 9 Yearly 46 46 Every 6 months 9 13 Every 3 months 6 9 Individual satisfaction Waterborne illnesses reported 0 Taste or odor problems 6 22 Satisfied with quality 97 240 In Canada, the lack of cistern usage data is highlighted by another review (Baird et al. 2013). There, underground cisterns in rural areas are commonly made from concrete, fiberglass, or plastic and are designed for long-term (3-5 month) storage of hauled truck water from another system, slow-producing wells, or collected precipitation. A Canadian cross-sectional postal survey of 246 residences with private water supplies conducted in May, 2004 revealed that 16% were supplied by cisterns (Jones et al. 2006). However, 38% of households using cistern water obtained water from two sources and 3% obtained water from 3 sources which included not only RHRW but also well-water stored in cisterns(Jones et al. 2006). An older study by Waller and Inman (1982) reported 23 households using collected rainwater (Lye 1992). One household used untreated roof water for drinking, six used a filter, and three used chlorine disinfectants. None of the 11 households using cisterns for their rainwater had cleaned the roof, five of the 11 tanks had not been cleaned, and no household diverted the first flush (Lye 1992, Waller and Inman 1982). For urban areas where access to municipal water systems is more likely, non- potable reuse applications may be of greater interest than use for drinking water. Gardeners, consumers, and members of the local community could come in contact with RHRW through touching or ingesting irrigated produce. Alaimo et al. (2008) conducted a cross-sectional phone survey of 766 Michigan adults regarding their fruit and vegetable intake from community gardens. Fifteen-percent reported that they or a member of their household had participated in a community garden over the past 12 months. Based on linear and logistic regression models, participants in this category were 1.4 times more likely to consume fruits and vegetables daily (p<0.001) and 3.5 times more likely to consume fruits and vegetables at least 5 times per day than those who did not participate in gardening(Alaimo et al. 2008). 241 9.1.3. Research gaps 9.1.3.1. Microbial ecology of RHRW systems and impact on water quality Communities of microorganisms play a role in RHRW quality through deposition of airborne material onto catchment (roof) surfaces, and during transport and storage. Brodie et al. (2007) describe the microbial diversity of urban aerosols, finding over 1,800 bacterial species via DNA microarray in aerosols from two sites in San Antonio and Austin, Texas, including organisms in the human pathogen- containing families Campylobacteraceae and Helicobacteraceae, indicating a potential for microbial deposition of human pathogens on RHRW catchment surfaces. Tree litter and animal fecal matter contribute to the deposition of contaminants on the catchment area and in rainwater containers (Lye 2009, TWDB 2005). The degree of contamination that occurs during collection and storage is highly dependent on several factors including antecedent conditions, locale, materials used for the construction of the RHRW system, characteristics of a precipitation event, and other hydrologic and meteorological factors. Conditions during the transport and storage of rainwater are a key area for future study. Rainwater storage tanks are enclosed and elevated, and therefore have inputs mostly limited to during rainfall events rather than from the soil environment or surface waters except in the case of animal or plant debris (Evans et al. 2009). However, the roof environment would receive incoming solar UV radiation (resulting in high intermittent temperatures), low nutrients, and desiccation, rendering it a harsh environment for microorganism survival. The inside of opaque tanks are devoid of 242 light and therefore would not provide favorable environments for photosynthetic bacteria or algae (Evans et al. 2009). Nutrient levels inside rainwater tanks are also low (oligotrophic), which would favor the survival of indigenous organisms over enteric (which could be human pathogenic) bacteria. It is unclear whether increased storage time would decrease water quality and warrant a low retention time as was noted for two studies for several parameters including turbidity, enterococcus, coliforms, nutrients, and zinc (Palla et al. 2011, van der Sterren et al. 2013), or if increased retention would improve the water quality of the system, potentially due to micoorganisms in biofilms extracting contaminants from the water (Coombes et al. 1999, Spinks et al. 2003). Drawdown inside the tank could expose the sides to oxygen, altering conditions for corrosion or other reactions (van der Sterren et al. 2013). Biological and chemical gradients exist in tanks, meaning that water at the surface is less contaminated than water at the bottom of the tank and it cannot be modeled as a homogenous system (Coombes 2002, van der Sterren et al. 2013). This highlights the importance of reporting sampling point locations for RHRW water quality determinations. Despite numerous observations of indicator and pathogenic microorganisms in the RHRW literature, the microbial ecology of domestic rainwater systems and storage tanks has been studied to a lesser extent. Although no North American/European examples are available, Evans et al. (2009) assessed the bacterial diversity within 22 domestic Australian rainwater storage tanks (n=83 samples) over a 2 year period, finding generally high diversity similar to soil, rainwater, and seawater (Madigan et al. 2012). The authors hypothesized that improvements in bacterial and chemical quality of RHRW can be due to naturally occurring physical processes of settling, but stable, biodiverse native bacterial ecosystems in the tanks can remove 243 nutrients and take part in bioremediation of contaminants and/or competitive exclusion of pathogenic microorganisms (Evans et al. 2009). A combination of culture and PCR analysis revealed 202 different species, with the majority (90%) from the phyla Proteobacteria (94% samples, average abundance >5000 CFU/ mL), Firmicutes (70% samples), Actinobacteria (27% samples), and Bacteroidetes (24% samples) (Evans et al. 2009). Coliform bacteria and species associated with fecal contamination were associated with <15% of the identified species and <1.5% total average abundance in rainwater tanks (Evans et al. 2009). These findings agreed with previous findings of complex microbial ecosystems on cistern tank surfaces from the US Virgin Islands by Isquith and Winters (1987) using traditional culture-based methods. These authors identified bacteria from the genera Pseudomonas, Flavobacterium Proteus, Bacillus, Achromobacter, and Serratia as well as total coliforms, fecal streptococcus, Salmonella sp., algae and protozoa (Isquith and Winters 1987). Alternatively, the organisms found in Evans et al. (2009) have been shown to degrade xenobiotics and metals (Madigan et al. 2012). The potential for bioremediation would be an argument for increased detention time, and post-tank disinfection, if applicable (Evans et al. 2006, van der Sterren et al. 2013). Further investigation into the impact of design and maintenance on the ecology and microorganism-mediated processes in rainwater tanks could provide better guidance as to their appropriate use and potential mitigation or control measures (Evans et al. 2009). It is known from studies of drinking water distribution systems and storage tanks that biofilms form on surfaces and can harbor human-pathogens (Lau and Ashbolt 2009b). Therefore, future work might investigate the microbial diversity of biofilms in rainwater tanks along with corresponding ability to harbor human 244 pathogens and their association with hydrologic/climate parameters, design, and maintenance schemes. Factors that influence the levels of pathogens in biofilms include turbidity, concentrations of assimilable organic carbon and biodegradable organic carbon in the water, biofilm particle surface properties, pipe material, flow rates, and water treatment scheme (Falkinham et al. 2001b, Långmark et al. 2005, Lin et al. 2012, Sibille et al. 1998). 9.1.3.2. Maintenance, treatment, and usage of RHRW Further investigation of RHRW usage, treatment, and system maintenance practices remains a gap to conducting a risk assessment. Although treatment schemes to purify harvested rainwater do exist, they can sometimes be complicated, and the most effective means of keeping harvesting rainwater clean is prevention of contamination through regular maintenance (Macomber 2004). An early study by Lye (1987) reported 16/30 of Kentucky households did never disinfected their water prior to drinking or did not use a filter (21/30), while 22/30 did not annually clean their systems or divert the first flush (2/30). More recently, Stump et al. (2012) surveyed 36 Texas households who were members of the Texas Rainwater Catchment Association, finding that eighty-one percent of households maintained their filters every 6 months, but 19% only did so every 6-12 months. A study of Bermuda rainwater tanks by Levesque et al. (2008) found that cleaning the tank the year before sampling resulted in a lower concentration of E. coli, however of the 102 households surveyed, 48% disinfected their tanks and only 40% did so on a regular basis. Due to presumed low maintenance and treatment rates among cistern owners, more information is needed to see which 245 applications of untreated rainwater predominate. When considering usages besides drinking, irrigation of urban community gardens with RHRW is of particular concern due to potential for ingestion of contaminants. 246 9.2. Screening paper supplemental 9.2.1. Supplemental figures and tables Table 9.4 Gene fragments used to construct plasmids for qPCR standards Opportunistic Target Amplicon Amplicon sequences pathogens Gene size (bp) Acanthamoeba 18S 65 CGACCAGCGATTAGGAGACGTTGAATACAAAACACCACC spp. rDNA ATCGGCGCGGTCGTCCTTGGCGTCGG Legionella spp. 23S 92 CCCATGAAGCCCGTTGAAGACTACGACGTTGATAGGCGA rRNA GGTGTGGAAGCGCAGTAATGTGTGAAGCTAACTCGTACT AATTGGCTGATTGT L. longbeachae mip 238 AGATGGGATGTCTGGTGCTCAATTGATTTTGACTGAAGA ACAAATGAAAGACGTTCTTAGTAAATTTCAGAAAGATCT GATGGCTAAGCGTAGTGCTGAGTTTAATAAAAAAGCAGA AGAAAACAAAGCAAAAGGTGACGCTTTCTTATCTGCTAA TAAGTCAAAACCTGGGATAGTAGTTTTACCAAGTGGTTT GCAGTATAAGATTATTGATGCCGGAACTGGTGCAAAACC AGGT L. pneumophila mip 78 AAGGCATGCAAGACGCTATGAGTGGCGCTCAATTGGCTT TAACCGAACAACAAATGAAAGACGTTCTTAACAAGTTTC Mycobacterium 16S 97 GGGTGAGTAACACGTGGGCAATCTGCCCTGCACTTCGGG avium rRNA ATAAGCCTGGGAAACTGGGTCTAATACCGGATAGGACCT CAAGACGCATGTCTTCTGG Mycobacterium 16S 100 GGGTGAGTAACACGTGGGCAATCTGCCCTGCACTTCGGG intracellulare rRNA ATAAGCCTGGGAAACTGGGTCTAATACCGGATAGGACCT TTAGGCGCATGTCTTTAGGTGG Pseudomonas regA 65 TGCTGGTGGCACAGGACATCCAGATGCTTTGCCTCAACG aeruginosa TCGACAATGAGGAACTGCACCAACAA 247 Table 9.5 qPCR primers, probes, and reaction mixtures used in quantitative PCR (qPCR) analysis Target Target Gene Primer and probe sequences (5’-3’)a Primer Probe Cycling parameters Amplicon Reference conc. conc. size (bp) (nM) (nM) Acanthamoeba 18S rDNA F: CGA CCA GCG ATT AGG AGA CG 240 240 10 min at 95°C, 45 cycles 65 (Rivière et al. spp. R: CCG ACG CCA AGG ACG AC of 15 s at 95°C, 15 s at 2006) P: FAM-TGA ATA CAA AAC ACC ACC ATC GGC GC-BHQ1 60°C, 15 s at 40°C Legionella spp. 23S rRNA F: CCC ATG AAG CCC GTT GAA 900 250 10 min at 95°C, 40 cycles 92 (Nazarian et al. R: ACA ATC AGC CAA TTA GTA CGA GTT AGC of 15 s at 95°C, 1 min at 2008) P: FAM-TCC ACA CCT CGC CTA TCA ACG TCG TAGT-TAMRA 60°C L. longbeachae mip F: AGA TGG GAT GTC TGG TGC TC 900 250 10 min at 95°C, 40 cycles 238 This work R: ACC TGG TTT TGC ACC AGT TC of 15 s at 95°C, 1 min at P: FAM-ACA AAG CAA AAG GTG ACG CT-BHQ1 60°C L. pneumophila mip F: AAA GGC ATG CAA GAC GCT ATG 900 250 10 min at 95°C, 40 cycles 78 (Nazarian et al. R: GAA ACT TGT TAA GAA CGT CTT TCA TTT G of 15 s at 95°C, 1 min at 2008) P: FAM-TGGCGCTCAATTGGCTTTAACCGA-TAMRA 60°C M. avium 16S rRNA F: GGG TGA GTA ACA CGT GTG CAA 900 200 50°C for 2 min, 95°C for 97 (Chern et al. 2015) R: CCA GAA GAC ATG CGT CGT GA 10 min, 40 cycles of 15 s P: FAM-TGC ACT TCG GGA TAA GCC TGG GAA A-TAMRA at 95°C, 1 min at 60°C M. intracellulare 16S rRNA F: GGG TGA GTA ACA CGT GTG CAA 900 200 50°C for 2 min, 95°C for 100 (Chern et al. 2015) R: CCA CCT AAA GAC ATG CGA CTA AA 10 min, 40 cycles of 15 s P: FAM-TGC ACT TCG GGA TAA GCC TGG GAA A-TAMRA at 95°C for, 1 min at 60°C P. aeruginosa regA F: TGC TGG TGG CAC AGG ACA T 1000 250 50°C for 2 min, 95°C for 65 (Shannon et al. R: TTG TTG GTG CAG TTC CTC ATT G 10 min, 45 cycles at a5 s 2007) P: FAM-CAG ATG CTT TGC CTC AA-BHQ1 at 95°C, 1 min at 60°C a F, Forward primer; R, reverse primer; P, probe; FAM, 6-carboxyflourescein; BHQ1, black hole quencher 1; TAMRA, 6-carboxytetramethylrhodamine 248 Table 9.6 Rainwater tank characteristics reported by Brisbane (n = 76 total survey respondents) and Currumbin (n = 45 total survey respondents) rainwater survey participants (one survey entry per tank). Note that multiple uses were reported for each tank and percentages were calculated using the total number of study participants that responded to the question in each location (n = 76 and n = 44, for Brisbane and Currumbin, respectively) Tank characteristic Location No. (%) Brisbane (n = 76) Currumbin (n = 45) Total (n = 121) Tank size (L) Nr = 73 Nr = 42 Nr=115 3,000 – 5,000 21 (29) 1 (2.4) 22 (19) >5,000 – 10,000 31 (43) 2 (4.8) 33 (29) >10,000 – 15,000 7 (10) 0 (0) 7 (6.1) >15,000 – 20,000 11 (15) 3 (7.1) 14 (12) >20,000 – 40,000 0 (0) 10 (24) 10 (8.7) >40,000 0 (0) 23 (55) 23 (20) Not sure 5 (7) 1 (2.4) 6 (5.2) Age of tank (years) (Mean, Range) Nr = 75 Nr = 42 Nr = 117 6.3 (0.5-20) 6.0 (0.5-20) 6.3 (0.5-20) Not sure 2 (2.7) 0 (0) 2 (1.7) Tank Material Nr = 75 Nr = 42 Nr=117 Polyethylene 51 (68) 6 (14) 57 (49) Galvanized 18 (24) 35 (83) 53 (45) Concrete 4 (5.3) 1 (2.4) 5 (4.3) Not sure 1 (1.3) 0 (0) 1 (0.8) Other 1 (1.3) 0 (0) 1 (0.8) Type of roof material Nr = 75 Nr = 42 Nr = 117 Metal 46 (61) 40 (95) 86 (74) Clay / concrete tile 29 (39) 0 (0) 29 (25) Green 0 (0) 1 (2.4) 1 (0.8) Not sure 0 (0) 1 (2.4) 1 (0.8) Presence of overhanging trees Nr = 75 Nr =42 Nr = 117 Yes 13 (17) 3 (7.1) 16 (14) No (5% or less) 62 (83) 39 (93) 101 (86) Evidence of wildlife droppings on roof Nr =73 Nr =40 Nr = 113 Yes 46 (63) 17 (43) 63 (56) No 27 (37) 23 (58) 50 (44) % of time in the shade morning/afternoon Nr =76 Nr = 41 Nr = 116 0 14 (18) 9 (22) 22 (19) 25 39 (51) 17 (42) 56 (48) 50 8 (11) 11 (27) 19 (16) 75 9 (12) 3 (7.3) 12 (10) 100 5 (7) 1 (2.4) 6 (5.2) *Sample numbers (Nr) indicate number of participants who responded to the survey in each location and indicated a response to the rainwater usage question 249 Table 9.7 Mean ± standard deviation (SD), range of Amplification efficiencies (E), Correlation coefficient (r2) and Slope of the standard curves for qPCR assays. Opportunistic Amplification efficiencies Correlation coefficient (r2) Slope pathogens (E) Acanthamoeba spp. 94.7 ± 2.11 (91.1, 96.6) 0.997 ± 0.003 (0.993, 0.999) -3.457 ± 0.058 (-3.556, -3.406) Legionella spp. 96.7 ± 2.81 (94.7, 100.8) 0.986 ± 0.018 (0.954, 0.997) -3.405 ± 0.070(-3.456, -3.304) L. longbeachae 93.4 ± 1.09 (92.7, 95.2) 0.992 ± 0.007 (0.982, 1.000) -3.492 ± 0.030 (-3.512, -3.442) L. pneumophila 94.9 ± 3.38 (93.0, 101) 0.990 ± 0.010 (0.973, 0.998) -3.454 ± 0.086 (-3.502, -3.301) M. avium 95.5 ± 0.84 (94.7, 96.9) 0.998 ± 0.002 (0.996, 0.999) -3.399 ± 0.021(-3.455, -3.399) M. inracellulare 93.5 ± 5.42 (86.2, 98.3) 0.987 ± 0.021 (0.949, 0.998) -3.495 ± 0.152(-3.703, -3.362) P. aeruginosa 95.7 ± 3.90 (92.0, 102) 0.994 ± 0.004 (0.988, 0.999) -3.431 ± 0.100 (-3.529, -3.286) E: Amplification efficiency = 10(1/−slope)/2. r2 : denotes the correlation coefficient of determination representing the proportion of variability accounted for by the linear model. 250 9.2.2. Recruitment of participants Figure 9.1 Recruitment letter 251 9.2.3. Survey administered to participants 252 253 254 255 256 Figure 9.2 Voluntary survey sent to rainwater study participants 257 9.2.4. Dissemination of study results to participants For the Currumbin Ecovillage, a presentation meeting was held on May 17, 2015 to provide information on the intent of the study and opportunities for questions and feedback from residents. Results and additional public health information were provided to each resident regarding their tank using the email text and attached results sheet below (Figure 9.3- 9.4). Participants were encouraged to contact the coordinators with questions and coordinators addressed all communications on an individual and confidential basis. 258 Dear [participant], Please find attached the results from our analysis of water samples from your rainwater tank. Most of these microorganisms are considered to be “opportunistic pathogens,” that is, they are commonly found in the environment (e.g. water and/or soil) but rarely cause illness in individuals in good health. They can, however, cause harm to those who have weakened immune systems, the very young, and the elderly. Legionella bacteria (Legionella spp.) were found in nearly all tanks. However, this is not surprising nor is it likely to be a cause for concern because they are a common group of bacteria found in soil, sediments and waters. Amongst this group of bacteria, Legionella pneumophila and Legionella longbeachae are the strains that cause most harm to humans. Legionella pneumophila can cause legionellosis which ranges from a mild febrile illness (Pontiac fever) to a serious form of pneumonia (Legionnaires’ disease) and Legionella longbeachae can cause fever, cough, chest pain, breathlessness, or diarrhoea. Legionella pneumophila was found in very few tanks (less than 4% of our total study of 134 tanks) whilst Legionellalongbeachae, the species that has been associated with some previous outbreaks related to potting soil, was not detected in any of the tanks tested. Please note that infection with Legionella pneumophila only occurs from inhaling sprays or mists, or when already ill people suck Legionella-contaminated water into their lungs. Infection does not result from healthy people consuming the water or from spreading person to person. Mycobacterium avium complex (MAC) is a group of bacteria that includes the two bacteria Mycobacterium avium and Mycobacterium intracellulare. MAC have caused pulmonary disease in some healthy individuals and cervical lymphadenitis in children, but the transmission to humans is not fully understood. MAC are known to occur in municipal water supplies, and there is ongoing research to try to establish whether these microorganisms pose any significant health risk. The probable exposure routes are through ingestion or inhalation of water or soil dusts. Acanthamoeba are a group of microorganisms (protozoans) that are common in the environment, including in tap water. Some types ofAcanthamoeba can cause eye infections (keratitis) in those who wear contact lenses and in more rare cases granulomatous amoebic encephalitis (GAE). Acanthamoeba can interact with Legionella bacteria in the environment and contribute to the growth and virulence of the Legionella, as well as increase the Legionella’s resistance to disinfection. We measured two types of bacteria that are associated with faecal contamination including Escherichia coli (E. coli) and Enterococci. There should not be any E. coli detected in drinking water according to the Australian Drinking Water guidelines (ADWG). Although there are no guideline values for enterococci for drinking water, no Enterococci should be present in samples of rainwater intended for drinking and for pool top up or recreational use the Australian and New Zealand Environment and Conservation Council (ANZECC) recreational water quality guidelines for fresh and marine waters is 35 entercocci/100 ml of water. We are also currently investigating factors that could be associated with the occurrence of microorganisms in tanks and conducting a health risk assessment known as a “Quantitative Microbial Risk Assessment (QMRA)” to decide which microorganisms and water use scenarios could present a health risk. We will send to you any future publications/reports from this research once it is available. Irrespective of the microorganisms detected in your tank, we strongly recommend proper maintenance of your tanks and, where necessary, treatment of the water in accordance with the enHealth Guidance on the use of rainwater tanks: 259 https://www.health.gov.au/internet/main/publishing.nsf/Content/0D71DB86E9DA7CF1CA25 7BF0001CBF2F/$File/enhealth-raintank.pdf. You can also find relevant health advice at the Queensland Government website: http://www.qld.gov.au/health/. If you require additional information about health risks of your rainwater tank, please call 13 HEALTH or 13 432584 and ask to be connected to “your local Public Health Unit.” If you have any queries regarding this study, please do not hesitate to contact me at [email] or [phone]. We sincerely appreciate your participation in the rainwater study. Additional Resources: http://conditions.health.qld.gov.au/HealthConditions/2/Infections-Parasites/6/Bacterial- Infections/914/Legionnaires-Disease NHMRC, NRMMC (2011). Australian Drinking Water Guidelines Paper 6 National Water Quality Management Strategy. National Health and Medical Research Council, Commonwealth of Australia, Canberra. Available at https://www.nhmrc.gov.au/_files_nhmrc/publications/attachments/eh52_australian_drinking_ water_guidelines_150527.pdf Figure 9.3 Email text used to disseminate results to rainwater study participants 260 Sample info Tank ID: Owner: Suburb: Size of the tank: End use: 1. Faecal indicator concentration/100 ml of water sample Indicator bacteria Colony forming unit (cfu)/100 ml Escherichia coli Enterococci Note: The number of Escherichia coli should be 0/100 ml in rainwater samples according to the Australian Drinking Water guidelines (ADWG). The presence of enterococci also indicates the presence of faecal pollution. However, there are no guideline values for enterococci for drinking water. For enterococci, the Australian and New Zealand Environment and Conservation Council (ANZECC) recreational water quality guidelines for fresh and marine waters is 35 entercocci/100 ml of water. Their number should also be 0 per 100 ml in rainwater samples. 2. Presence/absence of potential pathogens/L of water sample. Potential pathogens Presence/absence Legionella spp. Legionella pneumophila Legionella longbeachae Mycobacterium avium complex (“MAC”, subspecies hominissuis and M. intracellulare) Acanthamoeba spp. 261 Note: The presence of potential pathogens was determined using Polymerase Chain Reaction (PCR) method. The results only indicate the presence/absence of the tested pathogens based on the DNA analysis. Please note that these results do not provide information regarding the viability (i.e. whether the potential pathogens are live or dead). There should not be any pathogens present in drinking water. Please refer to the following page for a description of each pathogen. Significance of pathogens Legionella. The bacteria Legionella pneumophila can cause legionellosis which ranges from a mild febrile illness (Pontiac fever) to a potentially fatal form of pneumonia (Legionnaires’ disease) that can affect anyone, but primarily affects those who are susceptible due to age, illness, immunosuppression and other risk factors, such as smoking (WHO, 2007). Legionella bacteria are found in both natural and engineered aquatic environments. The bacteria grow best in warm water and cause infection when a person breathes a mist or vapor (small droplets of water in the air) containing the bacteria (CDC, 2015). The bacteria are not spread person to person. Legionella longbeachae is another bacteria in this group that can cause similar illness including fever, cough, chest pain, breathlessness, or diarrhoea. Both infections can be treated with antibiotics. Legionella longbeachae is typically not common in aquatic environments but the major source of human infection is considered to be commercial potting mixes and/or gardening and decomposing materials such as bark, sawdust, or composted animal manures (SA, 2015). Mycobacterium avium complex (MAC) Several different syndromes are caused by MAC, mostly in people whose immune systems are compromised (such as those with HIV/AIDS) (CDC, 2005). However, MAC causes pulmonary disease in healthy people and cervical lymphadenitis in children. Although MAC acquisition is an area of active research, probable routes of exposure (Falkinham 2013) are ingestion or inhalation of water or soil dusts. Acanthamoeba spp. are a group of microorganisms (protozoans) that are common in the environment, including in tap water. Some types of Acanthamoeba can cause eye infections (keratitis) in those who wear contact lenses and in more rare cases granulomatous amoebic encephalitis (GAE). Acanthamoeba can interact with Legionella bacteria in the environment and cause contribute to the growth and virulence of these bacteria, as well as their resistance to disinfection (USEPA, 2015). 262 References Falkinham III, J. O. (2013). Ecology of Nontuberculous Mycobacteria—Where Do Human Infections Come from? Seminars in respiratory and critical care medicine, Thieme Medical Publishers. Government of South Australia, 2015. Legionella longbeachae infection- including symptoms, treatment, and prevention. Available at http://www.sahealth.sa.gov.au US Centers for Disease Control and Prevention, 2005. Mycobacterium avium Complex. Available at http://www.cdc.gov/ncidod/dbmd/diseaseinfo/mycobacteriumavium_t.htm US Centers for Disease Control and Prevention. Legionella (Legionnaires’ Disease and Pontiac Fever): Causes & Transmission. Available at http://www.cdc.gov/legionella/about/causes-transmission.html US Environmental Protection Agency, 2015. Do you wear contact lenses? There’s something you should know. Available at http://water.epa.gov/action/advisories/acanthamoeba/index.cfm. World Health Organization, 2007. Legionella and the prevention of legionellosis. Available at http://www.who.int/water_sanitation_health/emerging/legionella.pdf Figure 9.4 Template for personalized results sheet provided to each rainwater study participant in screening and seasonal study 263 9.3. Seasonal rainwater paper supplemental Table 9.8 Opportunistic premise plumbing pathogens (OPPPs) in 134 tank water samples in phase 1. Tank water samples chosen for phase 2 indicated (*) Tank Concentration (Number per liter) Legionella spp. L. pneumophila Acanthamoeba spp. M. Avium P. aeruginosa M. intracellulare E. coli Enterococcus spp. 1 1.78E+05 23 2.59E+05 265 52 1.57E+04 266 81 1.68E+05 267 110 5.66E+05 268 Table 9.9: qPCR primers, probes, and reaction mixtures used in quantitative PCR (qPCR) analysis Target Target Primer and probe sequences (5’-3’)a Primer Probe Cycling parameters Amplicon Reference Gene conc. conc. size (bp) (nM) (nM) Acanthamoeba 18S F: CGA CCA GCG ATT AGG AGA CG 240 240 10 min at 95°C, 45 65 Rivière et al., spp. rDNA R: CCG ACG CCA AGG ACG AC cycles of 15 s at 2006 P: FAM-TGA ATA CAA AAC ACC ACC 95°C, 15 s at 60°C, ATC GGC GC-BHQ1 15 s at 40°C Legionella 23S rRNA F: CCC ATG AAG CCC GTT GAA 900 250 10 min at 95°C, 40 92 Nazarian et spp. R: ACA ATC AGC CAA TTA GTA CGA GTT cycles of 15 s at al., 2008 AGC 95°C, 1 min at 60°C P: FAM-TCC ACA CCT CGC CTA TCA ACG TCG TAGT-TAMRA L. mip F: AGA TGG GAT GTC TGG TGC TC 900 250 10 min at 95°C, 40 238 This work longbeachae R: ACC TGG TTT TGC ACC AGT TC cycles of 15 s at P: FAM-ACA AAG CAA AAG GTG ACG 95°C, 1 min at 60°C CT-BHQ1 L. mip F: AAA GGC ATG CAA GAC GCT ATG 900 250 10 min at 95°C, 40 78 Nazarian et pneumophila R: GAA ACT TGT TAA GAA CGT CTT TCA cycles of 15 s at al., 2008 TTT G 95°C, 1 min at 60°C P: FAM- TGGCGCTCAATTGGCTTTAACCGA- TAMRA M. avium 16S rRNA F: GGG TGA GTA ACA CGT GTG CAA 900 200 50°C for 2 min, 97 Chern et al., R: CCA GAA GAC ATG CGT CGT GA 95°C for 10 min, 40 2014 P: FAM-TGC ACT TCG GGA TAA GCC cycles of 15 s at TGG GAA A-TAMRA 95°C, 1 min at 60°C M. 16S rRNA F: GGG TGA GTA ACA CGT GTG CAA 900 200 50°C for 2 min, 100 Chern et al., intracellulare R: CCA CCT AAA GAC ATG CGA CTA AA 95°C for 10 min, 40 2014 P: FAM-TGC ACT TCG GGA TAA GCC cycles of 15 s at TGG GAA A-TAMRA 95°C for, 1 min at 60°C P. aeruginosa regA F: TGC TGG TGG CAC AGG ACA T 1000 250 50°C for 2 min, 65 Shannon et R: TTG TTG GTG CAG TTC CTC ATT G 95°C for 10 min, 45 al., 2007 P: FAM-CAG ATG CTT TGC CTC AA-BHQ1 cycles at a5 s at 95°C, 1 min at 60°C a F, Forward primer; R, reverse primer; P, probe; FAM, 6-carboxyflourescein; BHQ1, black hole quencher 1; TAMRA, 6-carboxytetramethylrhodamine 269 Table 9.10 Mean ± standard deviation (SD), range of Amplification efficiencies (E), Correlation coefficient (r2) and Slope of the standard curves for qPCR assays. Opportunistic Amplification efficiencies Correlation coefficient (r2) Slope pathogens (E) Acanthamoeba spp. 94.7 ± 2.11 (91.1, 96.6) 0.997 ± 0.003 (0.993, 0.999) -3.457 ± 0.058 (-3.556, -3.406) Legionella spp. 96.7 ± 2.81 (94.7, 100.8) 0.986 ± 0.018 (0.954, 0.997) -3.405 ± 0.070(-3.456, -3.304) L. longbeachae 93.4 ± 1.09 (92.7, 95.2) 0.992 ± 0.007 (0.982, 1.000) -3.492 ± 0.030 (-3.512, -3.442) L. pneumophila 94.9 ± 3.38 (93.0, 101) 0.990 ± 0.010 (0.973, 0.998) -3.454 ± 0.086 (-3.502, -3.301) M. avium 95.5 ± 0.84 (94.7, 96.9) 0.998 ± 0.002 (0.996, 0.999) -3.399 ± 0.021(-3.455, -3.399) M. inracellulare 93.5 ± 5.42 (86.2, 98.3) 0.987 ± 0.021 (0.949, 0.998) -3.495 ± 0.152(-3.703, -3.362) P. aeruginosa 95.7 ± 3.90 (92.0, 102) 0.994 ± 0.004 (0.988, 0.999) -3.431 ± 0.100 (-3.529, -3.286) E: Amplification efficiency = 10(1/−slope)/2. r2 : denotes the correlation coefficient of determination representing the proportion of variability accounted for by the linear model. 270 Table 9.11 List of meteorological stations closest to sampling sites 1-24 that provided rainfall only (Gauge A- L) or a complete set of meteorological data (Gauge 1- 5) Gauge Parameter Location BOM Station Number A Rainfall Brisbane Botanic Gardens Mt. Coot-Tha 040976 B Rainfall Alderley 040224 C Rainfall Mount Crosby TM 040818 D Rainfall Oxley 040463 E Rainfall Hilltop Gardens 040911 F Rainfall Brisbane Road Alert 040874 G, 1 Rainfall, Temperature, Brisbane 040913 RH H Rainfall Calamvale Alert 040784 I Rainfall Bellbird Park (Purser Road) 040985 J Rainfall Mango Hill 040986 K Rainfall Morayfield Alert 040979 L Rainfall Tallebudgera Guineas Creek Road 040196 2 Temperature, RH Amberley 040004 3 Temperature, RH Archerfield 040211 4 Temperature, RH Redcliffe 040958 5 Temperature, RH Coolangatta 040717 271 Table 9.12 Cochran Q test statistics and McNemar post-hoc testing results indicating which month(s) were significantly different. Significant results (p < 0.05 Cochran Q; McNemar post-hoc test p < 0.003, Bonferroni corrected value) shown in bold. Microorganism Cochran p- McNemar post- hoc (p-value) Q value Month Sep Oct Nov Feb Mar E. coli 14.125 0.015 Aug 1 0.754 0.424 0.146 0.013 Sep 0.754 0.289 0.07 0.002 Oct 0.754 0.424 0.039 Nov 0.774 0.146 Feb 0.125 Enterococcus 21.732 0.001 spp. Aug 1 0.125 0.004 0.013 0.039 Sep 0.109 0.002 0.003 0.012 Oct 0.344 0.302 0.549 Nov 1 1 Feb 0.727 Acanthamoeba 37.738 <0.001 spp. Aug 0.001 1 1 0.07 0.375 Sep 0.001 0.001 <0.001 <0.001 Oct 1 0.109 0.508 Nov 0.07 0.453 Feb 0.375 Legionella spp. 1.087 0.995 Aug 1 1 1 1 1 Sep 1 1 1 1 Oct 1 1 1 272 Nov 1 1 Feb 1 L. 10.600 0.06 pneumophila Aug 0.5 0.5 0.25 Sep 0.5 0.5 0.25 Oct 0.5 0.5 0.25 Nov 1 1 Feb 1 M. avium 34.000 <0.001 Aug 1 <0.001 1 0.688 0.289 Sep <0.001 1 0.453 0.508 Oct <0.001 <0.001 0.004 Nov 0.727 0.344 Feb 0.031 M. 11.132 0.049 intracellulare Aug 1 0.453 1 0.07 0.453 Sep 0.344 1 0.065 0.289 Oct 0.219 0.453 1 Nov 0.039 0.125 Feb 0.375 P. aeruginosa 72.683 <0.001 Aug 1 1 0.5 <0.001 0.039 Sep 1 1 <0.001 0.022 Oct 0.625 <0.001 0.039 Nov <0.001 0.002 Feb <0.001 273 Table 9.13 Friedman test statistics and Wilcoxon signed-rank post-hoc testing results indicating which month(s) were significantly different. Significant results (p < 0.05 Friedman test; Wilcoxon post-hoc test p < 0.003, Bonferroni corrected value) shown in bold. Microorganism Friedman test Wilcoxon signed-rank test Sep Oct Nov Feb Mar Χ2 p z p z p z p z P z p E. coli 22.331 <0.001 Aug -0.871 0.384 -0.236 0.814 -1.217 0.224 -1.63 0.103 -2.308 0.021 Sep -0.245 0.806 -1.964 0.05 -2.919 0.004 -3.243 0.001 Oct -1.51 0.131 -1.765 0.078 -3.333 0.001 Nov -1.133 0.257 -2.409 0.016 Feb -1.254 0.21 Enterococcus spp. 18.759 0.002 Aug -1.101 0.271 -1.112 0.266 -1.793 0.073 -2.065 0.039 -1.96 0.05 Sep -1.65 0.099 -1.794 0.073 -3.355 0.001 -2.356 0.018 Oct -0.052 0.959 -1.047 0.295 -0.403 0.687 Nov -1.547 0.122 -0.398 0.691 Feb -1.268 0.205 Acanthamoeba spp. 39.265 <0.001 Aug -1.279 0.201 -0.684 0.494 -1.364 0.172 -2.191 0.028 -2.346 0.019 Sep -2.226 0.026 -2.938 0.003 -4.08 <0.001 -3.959 <0.001 Oct -0.23 0.818 -1.807 0.071 -1.921 0.055 Nov -1.605 0.108 -2.014 0.044 Feb -0.687 0.492 Legionella spp. 65.312 <0.001 Aug -3.194 0.001 -3.543 <0.001 -1.771 0.076 -3.429 0.001 -4.045 <0.001 Sep -2.403 0.016 -1.399 0.162 -3.984 <0.001 -4.106 <0.001 Oct -2.616 0.009 -4.106 <0.001 -4.167 <0.001 Nov -3.924 <0.001 -4.197 <0.001 274 Feb -0.335 0.738 L. pneumophila 9.506 0.09 Aug <0.001 1 <0.001 1 -1.342 0.18 -1.342 0.18 -1.633 0.102 Sep <0.001 1 -1.342 0.18 -1.342 0.18 -1.633 0.102 Oct -1.342 0.18 -1.342 0.18 -1.633 0.102 Nov -1.069 0.285 -0.535 0.593 Feb -0.535 0.593 M. avium 52.945 <0.001 Aug -1.49 0.136 -3.171 0.002 -1.099 0.272 -2.971 0.003 -3.829 <0.001 Sep -3.829 <0.001 -0.227 0.82 -1.542 0.123 -0.451 0.652 Oct -2.8 0.005 -4.286 <0.001 -4.286 <0.001 Nov -1.893 0.058 -0.69 0.49 Feb -1.96 0.05 M. intracelluare 48.437 <0.001 Aug -0.224 0.823 -3.245 0.001 -2.777 0.005 -3.574 <0.001 -3.593 <0.001 Sep -3.696 <0.001 -3.061 0.002 -3.702 <0.001 -3.702 <0.001 Oct -3.593 <0.001 -1.602 0.109 -2.045 0.041 Nov -3.68 <0.001 -3.621 <0.001 Feb -1.083 0.279 P. aeruginosa 87.368 <0.001 Aug -1.604 0.109 -1.214 0.225 -0.535 0.593 -4.286 <0.001 -1.782 0.075 Sep -0.365 0.715 <0.001 1 -4.286 <0.001 -2.411 0.016 Oct -0.365 0.715 -4.286 <0.001 -2.471 0.013 Nov -4.286 <0.001 -2.045 0.041 Feb -4.143 <0.001 275 Table 9.14 Significant correlations between the presence or absence of FIB and pathogens with meteorological parameters, determined using binary logistic regression across all sampling events. Significant correlations (p < 0.016 and CI does not include 1) shown); see methods for discussion of correcting for multiple comparisons using a FDR approach. Rain Day 0 through -7 indicate days antecedent to the sampling event. Microorganism Predictor χ2 p-value Nagelkerke R Odds ratio (OR) 5th 95th square E. coli Rain Day -3 6.023 0.014 0.055 1.487 1.064 2.078 Min daily temp 8.243 0.004 0.074 1.105 1.029 1.188 Max daily temp 8.49 0.004 0.077 1.143 1.042 1.254 Average max 6.75 0.009 0.061 1.162 1.035 1.306 monthly temp Enterococcus spp. Min daily temp 14.3 <0.001 0.127 1.149 1.062 1.243 Max daily temp 11.77 0.001 0.106 1.173 1.066 1.291 Average max 13.49 <0.001 0.12 1.245 1.101 1.408 monthly temp Acanthamoeba spp. Rain Day -6 8.259 0.004 0.076 1.138 1.035 1.251 Min daily temp 20.62 <0.001 0.18 0.85 0.787 0.917 Max daily temp 16.08 <0.001 0.143 0.822 0.742 0.911 Average max 10.14 0.001 0.092 0.828 0.734 0.934 monthly temp Relative humidity 19.64 <0.001 0.172 0.914 0.875 0.956 9am Relative humidity 16.19 <0.001 0.144 0.95 0.925 0.976 3pm L. pneumophila Min daily temp 8.06 0.005 0.169 1.414 1.039 1.923 M. intracellulare Rain Day -5 7.028 0.008 0.064 1.1676 1.031 1.322 P. aeruginosa Time to last rain 6.601 0.010 0.063 1.137 1.02 1.267 event Last rainfall event 15.89 <0.001 0.148 0.813 0.724 0.913 depth Rain Day -5 13.39 <0.001 0.125 0.755 0.624 0.914 Min daily temp 25.46 <0.001 0.229 1.263 1.131 1.409 276 Max daily temp 51.60 <0.001 0.425 1.55 1.328 1.809 Average max 46.02 <0.001 0.386 1.692 1.395 2.052 monthly temp 277 Table 9.15 Correlations between microorganisms in roof-harvested rainwater with meteorological parameters. Significant values (p < 0.031) are bold-faced); see methods for discussion of correcting for multiple comparisons using a FDR approach. Rain Day 0 through -7 indicate days antecedent to the sampling event. Pathogen Time to Last Total Rainfal Rainfall Rainfall Rainfall Rainfall Rainfall Rainfall Rainfall Total Avg. max Daily Daily RH 9am RH (Kendall’s tau, last rainfall rain over l Day Day -1 Day -2 Day -3 Day -4 Day -5 Day -6 Day -7 monthly monthly min max 3pm P)a rain event past 7 0c rainfall temp temp temp event depth days E. coli -0.083 -0.087 0.065 0.067 0.093 0.039 0.126 0.079 -0.048 0.007 0.040 0.098 0.174 0.174 0.190 0.075 0.093 (0.097) (0.088) (0.204) (0.060) (0.004) (0.266) (<0.001) (0.075) (0.274) (0.878) (0.401) (0.054) (<0.001) (<0.001) (<0.001) (0.136) (0.066) Enterococcus 0.021 -0.052 0.038 0.051 0.016 -0.002 0.017 -0.014 <0.001 0.047 0.016 0.081 0.170 0.205 0.163 0.087 0.121 spp. (0.673) (0.295) (0.453) (0.137) (0.611) (0.961) (0.617) (0.752) 1 (0.318) (0.738) (0.108) (<0.001) (<0.001) (0.001) (0.080) (0.015) Acanthamoeba 0.012 -0.023 0.003 -0.079 -0.070 0.005 -0.029 0.027 0.095 0.102 0.029 -0.047 -0.163 -0.197 -0.188 -0.196 -0.170 spp. (0.798) (0.641) (0.945) (0.016) (0.015) (0.873) (0.372) (0.522) (0.024) (0.025) (0.529) (0.335) (<0.001) (<0.001) (<0.001) (<0.001) (0.001) Legionella spp. 0.009 0.074 0.024 -0.099 -0.132 -0.142 -0.105 -0.003 0.105 0.091 -0.012 -0.047 -0.239 -0.238 -0.255 -0.100 -0.061 (0.869) (0.185) (0.670) (0.020) (<0.001) (0.001) (0.013) (0.949) (0.035) (0.087) (0.830) (0.400) (<0.001) (<0.001) (<0.001) (0.074) (0.281) L. pneumophila 0.010 0.013 0.019 0.001 0.003 -0.001 0.003 -0.011 0.013 0.020 0.033 0.022 0.016 0.025 0.014 0.014 0.042 (0.543) (0.482) (0.297) (NC)b (NC)b (NC)b (NC)b (NC)b (NC)b (<0.001) (<0.001) (0.244) (0.373) (0.175) (0.440) (0.418) (0.021) M. avium 0.080 0.060 -0.085 -0.103 -0.051 -0.98 -0.022 0.046 0.020 -0.122 -0.197 -0.112 -0.062 -0.164 -0.064 -0.031 -0.038 (0.129) (0.259) (0.113) (0.008) (0.162) (0.011) (0.565) (0.337) (0.668) (0.015) (<0.001) (0.037) (0.245) (0.002) (0.232) (0.558) (0.481) M. -0.084 -0.032 0.026 -0.036 -0.042 -0.022 -0.038 0.006 0.106 0.107 0.038 -0.029 -0.243 -0.238 -0.257 -0.180 -0.143 intracellulare (0.117) (0.555) (0.628) (0.369) (0.258) (0.586) (0.337) (0.903) (0.033) (0.036) (0.453) (0.597) (<0.001) (<0.001) (<0.001) (0.001) (0.008) P. aeruginosa 0.142 -0.176 -0.180 -0.010 0.003 -0.002 0.024 -0.097 -0.154 -0.125 -0.097 -0.059 0.320 0.280 0.329 0.086 0.005 (0.001) (<0.001) (<0.001) (0.708) (0.901) (0.946) (0.364) (0.010) (<0.001) (0.003) (0.020) (0.196) (<0.001) (<0.001) (<0.001) (0.056) (0.908) a. Kendall’s tau is a nonparametric correlation coefficient measuring the monotonic association between y and x. Note that Kendall’s tau is typically approximately 0.15 lower than Spearman’s ρ and Pearson’s r for the same strength of association (Helsel and Hirsch 2002). b. A P value did not converge for some L. pneumophila estimates. c. Day refers to day in relation to sampling day, where 0 = on the sampling day. 278 Figure 9.5 Total monthly rainfall data for gauges closest to Brisbane (Gauge G) and the Gold Coast (Gauge L) Figure 9.6 Daily rainfall over six month sampling period for two gauges representative of Brisbane (Gauge G) and Gold Coast (Gauge L) sites. 279 Figure 9.7 Daily minimum and maximum temperatures observed at five gauges closest to the sampling sites where full meteorological datasets were available Figure 9.8 Daily relative humidity measurements taken at 9am and 3pm at five gauges closest to the sampling sites where full meteorological datasets were available. 280 9.4. MAC literature review Table 9.16 MAC quantification methods. A review of additional molecular (primarily non-quantitative) methodology is available in (Halstrom et al. 2015) and is recommended as a companion to this table. Method Target Reference Culture Mycobacterium spp. (Jenkins et al. 1982) qPCR Mycobacterium spp. (Adrados et al. 2011) M. avium (Chern et al. 2015, Feazel et al. 2009, Wilton and Cousins 1992) M. avium subsp. avium (Slana et al. 2010) M. avium subsp. hominissuis M. intracellulare (Chern et al. 2015, Wilton and Cousins 1992) IMS with qPCR M. avium subsp. paratuberculosis (Chern et al. 2015, Feazel et al. 2009, Slana et al. 2008, Vansnick et al. 2004, Whan et al. 2005a) FISH and epifluorescence microscopy M. avium (Lehtola et al. 2006) Acid-fast stain and light microscopy Mycobacterium spp. (Smithwick 1976) Auramine-Rhodamine Stain and Fluorescence Mycobacterium spp. (Smithwick 1976) Microscopy 281 Table 9.17 Summary of systematic literature review for MAC dose response (Q=Queried, I= Imported) Agricola Google PubMed Web of Science Engineering Medline (Ovid) Scholar Village 1st Keyword 2nd keyword (s) Q I Q I Q I Q I Q I Q I Mycobacterium avium Animal model 2 2 2 2 16 13 117 39 2 0 367 43 or M. avium Animal experiment 1 1 0 0 16 11 26 11 0 0 9 1 Dose response 2 2 0 0 116 57 83 18 2 1 328 52 Virulence 17 15 42 27 281 142 281 84 10 3 1471 184 Infectivity 1 1 2 1 12 8 13 5 65 16 301 14 Dosing 1 1 1 1 38 11 360 86 9 2 873 Experimental Infection 19 19 43 25 36 32 155 89 1 1 191 44 Mycobacterium Animal model 0 0 0 0 3 3 27 17 3 0 26 9 intracellulare Animal experiment 0 0 0 0 0 0 8 4 1 0 0 0 Dose response 0 0 0 0 13 7 18 9 2 0 0 0 Virulence 4 4 0 0 53 35 73 34 16 1 151 23 Infectivity 0 0 0 0 0 0 2 2 61 19 16 1 Dosing 0 0 0 0 3 3 85 32 8 0 54 6 Experimental Infection 2 2 4 2 2 2 27 15 1 0 12 5 Mycobacterium Animal model 0 0 0 0 0 0 0 0 0 0 0 0 chimaera Animal experiment 0 0 0 0 0 0 0 0 0 0 0 0 Dose response 0 0 0 0 0 0 0 0 0 0 0 0 Virulence 0 0 0 0 3 3 2 0 0 0 3 2 Infectivity 0 0 0 0 0 0 0 0 0 0 0 0 Dosing 0 0 0 0 1 1 1 1 0 0 2 0 Experimental Infection 0 0 0 0 0 0 2 0 0 0 0 0 282 Table 9.18 In vivo dose response datasets for human-relevant MAC species Ref Host Route Strain Endpoint Dose Positive Negative Total Tomioka et al. Female Intraveno M. avium, Human- Lung lesions (112 1.1e6 1 4 5 (1993a) C57BL/6 us pooled strains N- days) 1.8e6 0 4 4 Mice 254, N-289, N-302, 2.7e6 0 5 5 N-307, N-364 3.4e6 4 0 4 3.7e6 0 5 5 M. avium, Lung lesions 2.1e6 0 4 4 environment- 2.2e6 0 5 5 pooled N-395, N- 2.4e6 0 5 5 396, N-415, N-424 2.6e6 1 4 5 M. avium serovar 1 Lung lesions 1.3 e6 3 3 6 pooled strains N- 1.3 e6 6 0 6 289, N-356, N-357, 1.3 e6 6 0 6 N-364, N-445, 2.1 e6 6 0 6 N458, N-461 2.2 e6 6 0 6 3.0 e6 6 0 6 4.7 e6 6 0 6 M. avium serovar 9 Lung lesions 0.7 e6 2 4 6 pooled strains N- 2.9 e6 6 0 6 254, N-302 M. avium Serovar 8 Death 1.4e6 3 3 6 (human-derived)- 1.4 e6 0 6 6 pooled N-307, 1.6 e6 0 6 6 N339, N-361, N- 1.6 e6 0 6 6 444, N-463 3.8 e6 0 6 6 M. intracellulare, Lung lesions 1.1 e6 5 0 5 human- pooled N- 2.7 e6 4 0 4 237, N-256, N-260, 3.8 e6 4 0 4 N-283, N-292, N- 4.2 e6 4 0 4 310 4.3 e6 5 0 5 6.5 e6 5 0 5 M. intracellulare, Lung lesions 1.1e6 5 0 5 Environmental- 1.2e6 5 0 5 pooled N-409, N- 3.8e6 0 4 4 410, N-411, N-414 4.9e6 4 1 5 M. intracellulare, Death 1.6e6 2 4 6 Serovar 14 (human 1.3e6 6 0 6 –derived): N-244, 2.6e6 6 0 6 N-245, N-256, N- 1e6 1 5 6 291, N-299, N-349 1.9e6 6 0 6 5.1e6 0 6 6 M. intracellulare, Death 0.7 e6 6 0 6 Serovar 16 (human- 1.1 e6 6 0 6 derived): N-241, N- 1.6 e6 6 0 6 260, N-283, N-284, 1.7 e6 0 6 6 N-285, N-292, N- 2.0 e6 6 0 6 477 2.5 e6 6 0 6 3.3 e6 6 0 6 Tomioka et al. Female Intraveno M. avium N-445 Death (363 days) 1.3- 0 6 6 (1997) C57B1/6 us serovar 1 (human- 3.3e6 and derived) BALB/c Mice with MAC- susceptible Bcg genotype M. avium N-254 Death 1.3- 0 6 6 serovar 9(human- 3.3e6 derived) M. avium N-339 Death 1.3- 3 3 6 serovar 8(human- 3.3e6 derived) 283 M. intracellulare N- Death 1.3- 6 0 6 256 serovar 3.3e6 14(human-derived) M. avium N-258 Death 1.3- 6 0 6 serovar 16(human- 3.3e6 derived) Tuffley et al. Pigs Oral M. intracellulare Gross lesions in 0 0 3 3 (1973) type VI Mesenteric lymph 1.2e9 3 0 3 node (MLN) 1.5e9 1 2 3 2e9 3 0 3 Gross lesions in 0 0 3 3 Ileocaecal lymph node 1.2e9 1 2 3 (I/C) 1.5e9 0 3 3 2e9 0 3 3 Gross lesions in 0 0 3 3 mandibular lymph 1.2e9 0 3 3 node 1.5e9 1 2 3 2e9 0 3 3 Microscopic lesions in 0 2 1 3 Mesenteric lymph 1.2e9 3 0 3 node 1.5e9 2 1 3 2e9 3 0 3 Microscopic lesions in 0 3 0 3 Ileocaecal lymph node 1.2e9 2 1 3 1.5e9 3 0 3 2e9 0 3 3 Microscopic lesions in 0 0 3 3 mandibular lymph 1.2e9 0 3 3 node 1.5e9 1 2 3 2e9 0 3 3 Infection in MLN 0 0 3 3 1.2e9 3 0 3 1.5e9 2 1 3 2e9 3 0 3 Infection in I/C 0 0 3 3 1.2e9 3 0 3 1.5e9 1 2 3 2e9 3 0 3 Infection in 0 0 3 3 mandibular lymph 1.2e9 1 2 3 node (M) 1.5e9 2 1 3 2e9 2 1 3 Infection with other 0 2 1 3 serotypes in 1.2e9 0 3 3 mesenteric lymph 1.5e9 2 1 3 node 2e9 0 3 3 Infection with other 0 2 1 3 serotypes in ileocaecal 1.2e9 0 3 3 lymph node 1.5e9 0 3 3 2e9 0 3 3 Infection with other 0 2 2 4 serotypes in 1.2e9 1 2 3 mandibular lymph 1.5e9 0 3 3 node 2e9 1 2 3 Yangco et al. Male Intratrach MAC serotype 8 Infection-Culture 1e7 46 13 59 (1989) golden eal (Human-derived, positive 1e8 54 5 59 Syrian inoculatio AIDS patient) 5e8 52 8 60 hamsters n Infection- Lung 1e7 37 22 59 positive 1e8 42 17 59 5e8 40 20 60 Infection- 1e7 21 38 59 Disseminated 1e8 37 22 59 5e8 41 19 60 Infection-spleen 1e7 20 39 59 1e8 34 25 59 284 5e8 33 27 60 Infection- liver 1e7 5 54 59 1e8 8 51 59 5e8 15 45 60 Infection-bone 1e7 1 58 59 1e8 0 59 59 5e8 3 57 60 Collins et al. Calves M. intracellulare, Recovery of strains 8.2e7 2 1 3 (1983) strains 29 (pig), from feces 6.0e9 1 2 3 31(soil), 34 (soil), 1.7e10 0 2 2 37 (pig), 38 3.6e10 0 2 2 (human), 39 5.3e10 0 2 2 (human) Positive tissue samples 8.2e7 0 3 3 6.0e9 0 3 3 7.0e9 2 0 2 1.7e10 0 2 2 3.6e10 2 0 2 5.3e10 1 1 2 5.7e10 3 0 3 Death 8.2e7 0 3 3 6.0e9 0 3 3 7.0e9 2 0 2 1.7e10 0 2 2 3.6e10 0 2 2 5.3e10 0 2 2 5.7e10 2 1 3 M. avium strains 3 Recovery of strains 2.5e7 2 1 3 (goat), 5 (deer), 6 from feces 1.7e8 0 3 3 (pig), 15 (pigeon), 1.7e9 2 0 2 16 (deer), 17 (deer), 1.0e10 0 3 3 20 (goat), all non- 3.3e10 0 3 3 paratuberculosis Positive tissue samples 2.5e7 0 3 3 1.7e8 3 0 3 1.7e9 2 0 2 1.0e10 2 1 3 2.4e10 2 0 2 2.5e10 2 0 2 3.3e10 2 1 3 3.7e10 2 0 2 Death 2.5e7 0 3 3 1.7e8 0 3 3 1.7e9 0 2 2 1.0e10 0 3 3 2.4e10 2 0 2 2.5e10 1 1 2 3.3e10 0 3 3 3.7e10 1 1 2 Collins and B6D2 Intraveno M. avium strain 706 Death 1e6-1e8 0 10 10 Watson (1981) hybrid us mice Intraveno M. avium strain 724 Death 1e6-1e8 4 6 10 us Balian et al. Hamsters Oral Bacteremia (80 days) 1e8 3 69 72 (2003) Liver infection 1e8 19 53 72 (256days) Lung infection (80 1e8 5 67 62 days) Mesenteric lymph 1e8 45 27 72 node infection Agdestein et al. Pigs Oral M. avium Culture from fecal 5e9 4 1 5 (2012) hominissuis samples (6 weeks) Culture from 5e9 5 0 5 mandibular lymph nodes (6 weeks) Culture from tonsils (6 5e9 5 0 5 285 weeks) Culture from 5e9 1 4 5 bifurcationis sinn. Lymph nodes (6 weeks) Culture from jejunal 5e9 4 1 5 lymph nodes (6 weeks) Culture from ileocolici 5e9 3 2 5 lymph nodes (6 weeks) Culture from colici 5e9 2 3 5 lymph nodes (6 weeks) Culture from Jejunal 5e9 3 2 5 Peyer’s patches (6 weeks) Culture from Ileal 5e9 4 1 5 Peyer’s patches (6 weeks) Culture from 5e9 4 0 4 mandibular lymph nodes (12 weeks) Culture from 5e9 1 3 4 bifurcationis sinn. Lymph nodes (12 weeks) Culture from jejunal 5e9 4 0 4 lymph nodes (12 weeks) Culture from ileocolici 5e9 3 1 4 lymph nodes (12 weeks) Culture from colici 5e9 2 2 4 lymph nodes (12 weeks) IFN-γ assay (5 weeks) 5e9 5 5 10 Diarrhea 5e9 2 8 10 Granulomatous lesions 5e9 1 4 5 in tonsils (@ 6weeks post-inoculation) Granulomatous lesions 5e9 1 4 5 in liver (6 weeks) Granulomatous lesions 5e9 2 3 5 in jejunal lymph nodes (6 weeks) Granulomatous lesions 5e9 1 4 5 in Ileal Peyer’s patches Granulomatous lesions 5e9 4 0 4 in mandibular lymph nodes (12 weeks) Granulomatous lesions 5e9 2 2 4 in tonsils (12 weeks) Granulomatous lesions 5e9 1 3 4 in lung (12 weeks) Granulomatous lesions 5e9 3 1 4 in liver (12 weeks) Granulomatous lesions 5e9 4 0 4 in jejunal lymph nodes (12 weeks) Granulomatous lesions 5e9 4 0 4 in ileocolici lymph nodes (12 weeks) Granulomatous lesions 5e9 2 2 4 in colici lymph node (12 weeks) Granulomatous lesions 5e9 3 1 4 286 in Jejunal Peyer’s patches (12 weeks) Granulomatous lesions 5e9 3 1 4 in Ileal Peyer’s patches (12 weeks) Bertram et al. C57B16J Intraveno MAC 101 Death (at 4 weeks) 1e7 5 7 12 (1986) beige mice us (tail vein) Cross et al. Ferrets Oral M. avium Strain Lesions in tissues 5e6 1 8 9 (2000) 197/13933 (deer- derived) Culture from head 5e6 3 6 9 region Culture from 5e6 2 7 9 mesenteric lymph node Culture from thoracic 5e6 1 8 9 lymph node Hendricks et al. Rhesus IV M. avium 88415 Diarrhea 1e8 4 12 16 (2004) macaques (simian AIDS (SIV- isolate) inoculated and tested positive post- inoculation ) Wasting 1e8 16 0 16 Disseminated MAC 1e8 16 0 16 infection Rhesus IV M. avium 88415 Wasting 1e8 0 4 4 macaques- (simian AIDS SIV free isolate) Diarrhea 1e8 0 4 0 Disseminated MAC 1e8 0 4 0 infection Bermudez et al. Female Oral MAC 101 (serovar Disseminated 5e8 21 25 46 (1992) C57BL (gavage) – 1) (All strains infection- blood 5e4 12 34 46 bg+/bg+ five doses AIDS-patient- (bacteremia) (culture-4 beige mice on derived) weeks) alternate days Disseminated 5e8 46 0 46 infection- liver (culture- 4wks) Disseminated 5e8 46 0 46 infection- spleen (culture-4 wks) Disseminated 5e8 46 0 46 infection- appendix (culture- 4 wks) Death (4 weeks) 5e8 21 25 46 5e4 5 41 46 MAC 101 (serovar Disseminated 5e4 15 31 46 1) infection- blood (culture-8 wks) 100 (serovar 8) Disseminated 5e4 0 46 46 infection- blood (culture-8 wks) 104 (serovar 1) Disseminated 5e4 32 14 46 infection- blood (culture-8 wks) 116 (serovar 1) Disseminated 5e4 12 34 46 infection- blood (culture-8 wks) Inderlied et al. C57BL/6J- IV MAC 101 serotype Death 1e7 33 19 52 (1989a) bg beige 1 (human AIDS mice patient-derived) Inderlied et al. C57BL/6J- IV MAC 101 (derived Death 1e7 4 5 9 287 (1989b) bg beige from AIDS patient mice with disseminated MAC disease) Ji et al. (1994) C57BL/6J Intraveno MAC 101 Death (4 weeks) 1e6.7 6 78 84 bg/bg beige us 1e7.91 16 33 49 mice Death (12 weeks) 1e6.94 7 6 13 MAC Lpr 1e7.55 8 9 17 MAC MO1 1e7.13 10 20 30 (All strains derived from human AIDS patient with disseminated MAC infection) Ji et al. (1996) Female Intraveno MAC 101 (passed Death (105 days) 1.87e7 31 11 42 C57BL/6J us through beige mice) bg/bg mice Kansal et al. Beige mice IV MAC 101 SmT Death (20, 30, and 40 5e7 2 18 20 (1998) morphotype days- remained unchanged) MAC 101 RgT Death (20 days) 5e7 2 18 20 morphotype Death (30 days) 5e7 10 10 20 Death (40 days) 5e7 14 6 20 Kanyok et al. C57BL- Intraveno MAC 101 (serotype Death (8 weeks) 1.16e8 5 2 7 (1994) 6/bg/bg us 1) beige mice Ledwon et al. Birds Intramusc M. intracellulare Lesions in liver (70 5e5 1 5 6 (2008) (Melopsitta ular strain 13950 days) cus undulates or budgerigar s) Lesions in intestines 1 5 6 (70 days) Lesions in pectoral 1 5 6 muscles (70 days) Present in feces by 1 5 6 microscopy Lorian et al. Young Intraperito M. intracellulare Lung culture 1.6e8 9 (1 mo.) 1 10 (1975) guinea pigs neal Boone strain 5 (3 mo.) 5 injection Spleen culture 1.6e8 10 (1 mo.) 0 10 2 (3 mo.) 8 Liver culture 1.6e8 8 (1 mo.) 2 10 4 (3 mo.) 6 Peritoneal fluid culture 1.6e8 8 (1 mo.) 2 10 2 (3 mo.) 8 M. intracellulare- Lung culture 1.1e8 8 (1 mo.) 2 10 unclassified strain 3 (3 mo.) 7 Spleen culture 1.1e8 7 (1 mo.) 3 10 0 (3 mo.) 10 Liver culture 1.1e8 6 (1 mo.) 4 10 4 (3 mo.) 6 Peritoneal fluid culture 1.1e8 7 (1 mo.) 3 10 6 (3 mo.) 4 M. intracellulare- Lung culture 3.9e7 9 (1 mo.) 1 10 Davis strain 4 (3 mo.) 6 Spleen culture 3.9e7 10 (1 mo.) 0 10 1 (3 mo.) 9 Liver culture 3.9e7 10 (1 mo.) 0 10 6 (3 mo.) 4 Peritoneal fluid culture 3.9e7 10 (1 mo.) 0 10 7 (3 mo.) 3 M. intracellulare Lung culture 7.1e7 9 (1 mo.) 1 10 Yandle-Wilson 1 (3 mo.) 9 strain Spleen culture 7.1e7 8 (1 mo.) 2 10 288 1 (3 mo.) 9 Liver culture 7.1e7 8 (1 mo.) 2 10 2 (3 mo.) 8 Peritoneal fluid culture 7.1e7 10 (1 mo.) 0 10 3 (3 mo.) 7 Adult Intraperito M. intracellulare Lung culture 1.6e8 4 (1 mo.) 6 10 guinea pigs neal Boone strain 1 (3 mo.) 9 injection Spleen culture 1.6e8 1 (1 mo.) 9 10 0 (3 mo.) 10 Liver culture 1.6e8 3 (1 mo.) 7 10 2 (3 mo.) 8 Peritoneal fluid culture 1.6e8 5 (1 mo.) 5 10 0 (3 mo.) 10 M. intracellulare- Lung culture 1.1e8 3 (1 mo.) 7 10 unclassified strain 0 (3 mo.) 10 Spleen culture 1.1e8 4 (1 mo.) 6 10 0 (3 mo.) 10 Liver culture 1.1e8 3 (1 mo.) 7 10 1 (3 mo.) 9 Peritoneal fluid culture 1.1e8 3 (1 mo.) 7 10 3 (3 mo.) 7 M. intracellulare- Lung culture 3.9e7 2 (1 mo.) 8 10 Davis strain 5 (3 mo.) 5 Spleen culture 3.9e7 3 (1 mo.) 7 10 4(3 mo.) 6 Liver culture 3.9e7 1(1 mo.) 9 10 3(3 mo.) 7 Peritoneal fluid culture 3.9e7 4(1 mo.) 6 10 3(3 mo.) 7 M. intracellulare Lung culture 7.1e7 1(1 mo.) 9 10 Yandle-Wilson 0(3 mo.) 10 strain Spleen culture 7.1e7 2(1 mo.) 8 10 1(3 mo.) 9 Liver culture 7.1e7 1(1 mo.) 9 10 3(3 mo.) 7 Peritoneal fluid culture 7.1e7 1(1 mo.) 9 10 0(3 mo.) 10 Lounis et al. Female IV MAC 101 Death 2.66e7 3 (45 d) 21 24 (1997) C57B1/6J/ bgi/bgi beige mice Matsuyama et Mice Intratrach M. avium Death (10 months) 1e7 6 14 20 al. (2014) eal hominissuis McGavin et al. Calves Intraderm M. intracellulare Death (79 days) 1e8 1 4 5 (1977) al serotype Davis (calf derived) Disseminated 1e8 3 2 5 granulomas Mehta (1996) Beige mice IV MAC 101 (human- Death (60 days) 1e8 6 14 20 derived) 5e7 6 14 20 2e7 0 20 20 1e7 0 20 20 1e6 0 20 20 (30 days) 1e8 6 14 20 5e7 4 16 20 2e7 0 20 20 1e7 0 20 20 1e6 0 20 20 Nakamura et al. Pigs Oral M. intracellulare Lesions- tonsils (12 2e7 2 0 2 (1984) serotype 8(T-146) weeks) 2e9 2 0 2 Lesions- mandibular 2e7 1 1 2 (12 weeks) 2e9 2 0 2 Lesions- jejunal (12 2e7 1 1 2 weeks) 2e9 2 0 2 Lesions-ileocolic (12 2e7 1 1 2 weeks) 2e9 0 2 2 289 Lesions- superficial 2e7 0 2 2 inguinal lymph node 2e9 1 1 2 (12 weeks) Brown et al. Cyclospori Oral M. avium strain Bacteremia (160 days) 1e9 1 2 3 (1991) ne-treated SK059, serotype 4 Sprague- (bronchoalveoloar Dawley lavage fluid of rats AIDS patient with cavitary poulmonary disease and granuloma on biopsy and later developed disseminated infection) Intraveno SK059 Bacteremia (36 days) 1e8 5 0 5 us (cyclosporine+ low dose steroids, 3 per group, no significant difference) Oral SK059 Death (140 days) 1e9 1 3 4 Oral Strain SK005, Bacteremia (160 days) 1e9 4 0 4 serotype unknown (human AIDS patient synovial fluid-derived) Oral SK005 Bacteremia (120 days) 1e9 3 3 6 1e6 0 5 5 Oral SK005 Present in stool (4 1e9 6 0 6 months) 1e6 0 6 6 SK090 1e9 0 5 5 Intraveno M. avium strain Bacteremia (36 days) 1e8 2 4 6 us SK150, serotype 8 (human blood AIDS patient derived) (cyclosporine+ low dose steroids, 3 per group, no significant difference) Oral SK090, serotype 4 Bacteremia (120 days) 1e9 0 5 5 (from non-AIDS patient sputum with lung carcinoma Waggie et al. C57BL/6N Oral (trial M. avium- Positive lung culture- 1.7e7 2 35 37 (1983) mice from 3) intracellulare adult females B6C3F1 hybrid production colony Oral Positive lung culture- 1.7e7 42 juvenile females Oral Gross lesions- adult 1.7e7 1 36 37 females Oral Gross lesions- juvenile 1.7e7 0 36 36 females Oral Pulmonary 1.7e7 2 35 37 microgranulomas- adult females Oral Pulmonary 1.7e7 2 34 36 microgranulomas- juvenile females Subcutane Positive lung culture- 8.5e7 8 34 42 ous adult females Subcutane Positive lung culture- 8.5e7 6 30 36 ous juvenile females Subcutane Gross lesions- adult 8.5e7 10 32 42 ous females 290 Subcutane Gross lesions- juvenile 8.5e7 1 35 36 ous females Subcutane Pulmonary 8.5e7 14 28 42 ous microgranulomas- adult females Subcutane Pulmonary 8.5e7 9 27 36 ous microgranulomas- juvenile females Ellsworth et al. Pigs Oral M. avium (pig- Culture from feces 1e8 5 0 5 (1980) derived strains of serotypes 1 and 2- animals from all serotypes pooled) (70-79 days from tissue culture and intestine swabs) Culture from Cervical 1e8 5 0 5 lymph node tissue Culture from Tonsil 1e8 5 0 5 Culture from 1e8 2 3 5 Mediastinal lymph nodes Culture from Spleen 1e8 2 3 5 Culture from Pancreas 1e8 0 5 5 Culture from Liver 1e8 1 4 5 Culture from Kidney 1e8 1 4 5 Culture from 1e8 2 3 5 Gastrohepatic lymph nodes Culture from Pelvic 1e8 1 4 5 lymph nodes Culture from 1e8 5 0 5 Mesenteric lymph nodes Culture from 1e8 2 3 5 Duodenal aggregated lymphatic follicles Culture from Jejunal 1e8 4 1 5 aggregated lymphatic follicles Culture from Ileal 1e8 3 2 5 aggregated lymphatic follicles Culture from Caudal 1e8 0 5 5 part of ileum Culture from 1e8 1 4 5 Ileocecal valve Culture from Cecum 1e8 1 4 5 Culture from Spiral 1e8 0 5 5 colon Lesions- cervical 1e8 4 1 5 lymph nodes (70-79 days) Lesions- tonsils 1e8 4 1 5 Lesions- mediastinal 1e8 0 5 5 lymph nodes Lesions- spleen 1e8 0 5 5 Lesions- pancreas 1e8 0 5 5 Lesions- liver 1e8 3 2 5 Lesions- kidney 1e8 0 5 5 Lesions- gastrohepatic 1e8 0 5 5 lymph nodes Lesions- pelvic lymph 1e8 0 5 5 nodes Lesions- mesenteric 1e8 5 0 5 lymph nodes Lesions- duodenal 1e8 0 5 5 aggregated lymphatic 291 follicles Lesions- jejunal 1e8 1 4 5 aggregated lymphatic follicles Lesions- ileal 1e8 3 2 5 aggregated lymphatic follicles Lesions- caudal part of 1e8 3 2 5 ileum Lesions- ileocecal 1e8 4 1 5 valve Lesions- cecum 1e8 2 3 5 Lesions- spiral colon 1e8 0 5 5 Slana et al. Pigs Oral M. avium Infection- all locations 5.46e7 8 2 10 (2010) hominissuis (T3488) Lesions- all locations 7 3 10 Lymph node lesions- 1 9 10 liver Lymph node lesions - 1 9 10 duodenal (beginning) Lymph node lesions - 1 9 10 jejunal (beginning) Lymph node lesions - 1 9 10 jejunal (middle) Lymph node lesions - 1 9 10 ileal (beginning) Lymph node lesions - 3 7 10 ileocaecal valve Mucosa lesions- 1 9 10 jejunal (beginning) Mucosa lesions- 1 9 10 jejunal (end) Mucosa lesions- ileal 1 9 10 (beginning) Lymph node culture- 1 9 10 ileal node (beginning) qPCR isolation- lymph 1 9 10 node- submandibular qPCR isolation- lymph 2 8 10 node- tracheobronchial Jejunal (middle) 2 8 10 Mucosa- duodenal 2 8 10 (beginning) Mucosa-jejunal 1 9 10 (middle) Muscles- musculus 2 8 10 gluteus Jorgensen Pigs Intraveno M. avium serotype 2 Lesions- mandibular 7.8e4 0 2 2 (1977) us strain SSC 1323 (same results for 7.8e5 0 2 2 (pig origin) lesions in other lymph 7.8e6 0 2 2 nodes: parotideus, 7.8e7 1 1 2 Lymph node lesions retropharyng. lat., 3.9e8 2 0 2 retropharyng. med., cervicalis sup. dors., cervicalis sup. ventr., subiliacus, popliteus, inguinalis prof., inguinalis superf.) Lesions- mesenterieus 7.8e4 2 0 2 7.8e5 2 0 2 7.8e6 1 1 2 7.8e7 1 1 2 3.9e8 2 0 2 Lesions- 7.8e4 1 1 2 tracheobronchalis 7.8e5 2 0 2 7.8e6 1 1 2 7.8e7 1 1 2 292 3.9e8 2 0 2 Lesions- hepatiens 7.8e4 2 0 2 7.8e5 2 0 2 7.8e6 1 1 2 7.8e7 2 0 2 3.9e8 2 0 2 Lesions- spleen 7.8e4 0 2 2 7.8e5 0 2 2 7.8e6 0 2 2 7.8e7 1 1 2 3.9e8 2 0 2 Lesions- liver 7.8e4 0 2 2 7.8e5 0 2 2 7.8e6 1 1 2 7.8e7 2 0 2 3.9e8 2 0 2 Lesions-lung 7.8e4 0 2 2 7.8e5 0 2 2 7.8e6 2 0 2 7.8e7 2 0 2 3.9e8 2 0 2 Lesions- kidney 7.8e4 0 2 2 7.8e5 0 2 2 7.8e6 1 1 2 7.8e7 2 0 2 3.9e8 1 1 2 Lesions- myocardium 7.8e4 0 2 2 7.8e5 0 2 2 7.8e6 0 2 2 7.8e7 0 2 2 3.9e8 0 2 2 Lesions- Musc. Lon. 7.8e4 0 2 2 Dorsi 7.8e5 0 2 2 7.8e6 0 2 2 7.8e7 0 2 2 3.9e8 1 1 2 Lesions- Musc. 7.8e4 0 2 2 Triceps brachii (same 7.8e5 0 2 2 for Musc. Biceps 7.8e6 0 2 2 femoris and Musc. 7.8e7 0 2 2 Psoas major) 3.9e8 0 2 2 Lesions- Tonsil 7.8e4 0 2 2 7.8e5 0 2 2 7.8e6 0 2 2 7.8e7 1 1 2 3.9e8 2 0 2 Lesions- Intestinal 7.8e4 0 2 2 mucosa (peyer patch) 7.8e5 0 2 2 7.8e6 1 1 2 7.8e7 2 0 2 3.9e8 2 0 2 Culture from lymph Culture-Lymph node 7.8e4 0 2 2 nodes mandibularis 7.8e5 1 1 2 7.8e6 0 2 2 7.8e7 2 0 2 3.9e8 2 0 2 Parotideus 7.8e4 0 2 2 7.8e5 0 2 2 7.8e6 0 2 2 7.8e7 2 0 2 3.9e8 2 0 2 Retropharyng. lat. 7.8e4 0 2 2 7.8e5 0 2 2 7.8e6 1 1 2 7.8e7 1 1 2 3.9e8 2 0 2 Retropharyng. med. 7.8e4 0 2 2 7.8e5 1 1 2 7.8e6 0 2 2 293 7.8e7 1 1 2 3.9e8 2 0 2 Cervicalis sup. 7.8e4 0 2 2 dors.(same for 7.8e5 0 2 2 Cervicalis sup. ventr., 7.8e6 0 2 2 subiliacus) 7.8e7 1 1 2 3.9e8 2 0 2 Popliteus 7.8e4 0 2 2 7.8e5 0 2 2 7.8e6 0 2 2 7.8e7 0 2 2 3.9e8 2 0 2 Ingninalis prof. (same 7.8e4 0 2 2 for Ingninalis superf.) 7.8e5 0 2 2 7.8e6 0 2 2 7.8e7 1 1 2 3.9e8 2 0 2 Mesentericus 7.8e4 0 2 2 7.8e5 2 0 2 7.8e6 1 1 2 7.8e7 2 0 2 3.9e8 2 0 2 Tracheobronchial. 7.8e4 0 2 2 Sin.(same for 7.8e5 2 0 2 hepaticus) 7.8e6 2 0 2 7.8e7 1 1 2 3.9e8 2 0 2 Spleen 7.8e4 0 2 2 7.8e5 0 2 2 7.8e6 0 2 2 7.8e7 2 0 2 3.9e8 2 0 2 Liver 7.8e4 0 2 2 7.8e5 0 2 2 7.8e6 0 2 2 7.8e7 1 1 2 3.9e8 2 0 2 Lung 7.8e4 0 2 2 7.8e5 2 0 2 7.8e6 1 1 2 7.8e7 2 0 2 3.9e8 2 0 2 Kidney (same for 7.8e4 0 2 2 myocardium, musc. 7.8e5 0 2 2 Long. Dorsi., musc. 7.8e6 0 2 2 Triceps brachii., 7.8e7 1 1 2 musc.) 3.9e8 2 0 2 Musc. Biceps femoris, 7.8e4 1 1 2 joint, tonsil, bone 7.8e5 0 2 2 marrow, intestinal 7.8e6 0 2 2 mucosa (peyer patch) 7.8e7 1 1 2 3.9e8 2 0 2 Musc. Psoas major 7.8e4 0 2 2 7.8e5 2 0 2 7.8e6 0 2 2 7.8e7 0 2 2 3.9e8 2 0 2 Gangadharam et C57B1/6J IV M. intracellulare Death (4 weeks) 1e7 8 12 20 al. (1983) bg/bg beige 571-8 mice C57B1/6J- 0 20 20 +/+) control mice 294 Table 9.19 Environmental sources where MAC species were identified. For additional discussion of MAC isolated from animals, animal foods, animal and vegetable food products, and agricultural wastes, the reader is referred to (Pavlik et al. 2009c). Media Environmental source Reference type Air Aerosols (Angenent et al. 2005, Gruft et al. 1981, Kirschner et al. 1999, Parker et al. 1983, Rautiala et al. 2004, Rhodes et al. 2014, Thomson et al. 2013a, Wendt et al. 1980) Water Aquaculture ponds (Klanicova et al. 2014, Makovcova et al. 2014) Aquarium water (Goslee and Wolinsky 1976, Yajko et al. 1995) Biofilms/sediments (Angenent et al. 2005, Chern et al. 2015, Falkinham III et al. 2001, Falkinham 2010, Falkinham 2011, (environmental) Iakhiaeva et al. 2016, Krizova et al. 2010, Matlova et al. 2003, Norton and LeChevallier 2000a, Ovrutsky et al. 2013, Pickup et al. 2006, Pryor et al. 2004b, Richards et al. 2015, Tichenor et al. 2012, Wallace et al. 2013, Whiley et al. 2015a) Bottled water (Yajko et al. 1995) Commercial (Al Sulami et al. 2012, Aronson et al. 1999, Band et al. 1982, Briancesco et al. 2014, Bukh and Roslev building/household premise 2014, Chilima et al. 2006, Donohue et al. 2015, Du Moulin and Stottmeier 1986, Falkinham III et al. plumbing & tap water 2008, Falkinham 2011, Fleischman et al. 1982, Glover et al. 1994, Goslee and Wolinsky 1976, Havelaar et al. 1985, Hilborn et al. 2006, Hilborn et al. 2008, Iakhiaeva et al. 2016, Ichijo et al. 2014, Iwamoto et al. 2012, Kamala et al. 1994b, Klanicova et al. 2013, Le Dantec et al. 2002b, Mansfield and Lackner 1997, Minomo et al. 2015, Montecalvo et al. 1994, Nishiuchi et al. 2007, Nishiuchi et al. 2009, Pelletier et al. 1988, Peters et al. 1995, Rhodes et al. 2014, Richards et al. 2015, Ristola et al. 2015, Stine et al. 1987, Thomson et al. 2013a, Tichenor et al. 2012, von Reyn et al. 2002, von Reyn et al. 1993b, Wallace et al. 2013, Whiley et al. 2015a, Yajko et al. 1995) Green building premise (Rhoads et al. 2016) plumbing & tap water Cooling tower water (Black and Berk 2003, Pagnier et al. 2009, Torvinen et al. 2014) Dam water and sediment (Whittington et al. 2003) Drinking water distribution (Covert et al. 1999, Falkinham III et al. 2001, Falkinham 2011, Hilborn et al. 2006, Krizova et al. 2010, systems Le Dantec et al. 2002b, Matlova et al. 2003, Powell and Steadham 1981, Pryor et al. 2004b, Thomson et al. 2013b, Wang et al. 2012a, Whiley et al. 2014a) Drinking water treatment (Hilborn et al. 2006, Klanicova et al. 2013, Le Dantec et al. 2002a, Pickup et al. 2006, Thomson et al. plant 2013b) Fountains (Du Moulin and Stottmeier 1986, Goslee and Wolinsky 1976) Hospital premise plumbing (Álvarez et al. 2008, Aronson et al. 1999, Bennett et al. 1994, Briancesco et al. 2014, Carson et al. 1988a, & tap water Carson et al. 1988b, Crago et al. 2014, du Moulin et al. 1988, Eaton et al. 1995b, Falkinham 2010, Glover et al. 1994, Graham et al. 1988, Gross et al. 1975, Mills 1972, Ovrutsky et al. 2013, Pelletier et al. 1988, Peters et al. 1995, Ristola et al. 2015, Ristola et al. 1999, Tobin-D’Angelo et al. 2004a, von Reyn et al. 1994) Hot tubs, spas, public baths, (Aksamit 2003, Angenent et al. 2005, Cappelluti et al. 2003, Embil et al. 1997, Fjällbrant et al. 2013, 295 footbaths, and pools Glazer et al. 2008, Glazer et al. 2007, Hanak et al. 2006, Havelaar et al. 1985, Kahana et al. 1997, Khoor et al. 2001, Koschel 2006, Lumb et al. 2004, Mangione et al. 2001, Marchetti et al. 2004, Marras et al. 2005, Moraga-McHaley et al. 2013, O'Neil et al. 2006, Pham et al. 2003, Rickman et al. 2002, Saito and Tsukamura 1976, Singer and Rodda 1965, Sood et al. 2007, Sugita et al. 2000, Thomson et al. 2013a, Travaline and Kelsen 2003, Urabe and Saito 1962, Vugia et al. 2005) Groundwater (Chilima et al. 2006, Falkinham 2011, Goslee and Wolinsky 1976, Kamala et al. 1994b, Le Dantec et al. 2002a, Ristola et al. 1999, von Reyn et al. 1993b) Ice machine (Du Moulin and Stottmeier 1986, Goslee and Wolinsky 1976) Marine waters/ beaches (Brooks et al. 1984b, Falkinham III et al. 1980, George et al. 1980, Goslee and Wolinsky 1976, Gruft et al. 1981, Singer and Rodda 1965) Point of use water filter (Rodgers et al. 1999) Direct rainfall (Wendt et al. 1980) Harvested rainwater (Hamilton et al. 2016, Lumb et al. 2004, Singer and Rodda 1965, Thomson et al. 2013a, Tuffley 1980) Reclaimed water distribution (Whiley et al. 2015b) system Reservoir or dam (Aronson et al. 1999, Glover et al. 1994, Goslee and Wolinsky 1976, Kamala et al. 1994a, Makovcova et al. 2014, Pelletier et al. 1988, Pickup et al. 2005, Pickup et al. 2006, Thomson et al. 2013b, Whan et al. 2005b) Reservoir or dam sediments (Klanicova et al. 2013) Space station “Mir” water (Kawamura et al. 2001) system Surface water (freshwater) (Bannalikar and Verma 2007, Bland et al. 2005, Brooks et al. 1984b, Covert et al. 1999, Eaton et al. 1995b, Falkinham 2011, George et al. 1980, Goslee and Wolinsky 1976, Gruft et al. 1981, Hilborn et al. 2006, Kamala et al. 1994c, Katila et al. 1995, Kazda 1973, Le Dantec et al. 2002a, Lee et al. 2008, Makovcova et al. 2014, Norby et al. 2007, Parashar et al. 2004, Parker et al. 1983, Pearson et al. 1977, Pelletier et al. 1988, Pickup et al. 2005, Pickup et al. 2006, Rhodes et al. 2013, Ristola et al. 1999, von Reyn et al. 1993b, Whan et al. 2005b, Whittington et al. 2003) Surface water (freshwater) (Pickup et al. 2005, Pickup et al. 2006) sediments Acidic, brown-water (Kirschner et al. 1992) swamps Wastewater (Jones et al. 1981, Pickup et al. 2005, Pickup et al. 2006) Soil, Boreal forest soils and peat (Bauer et al. 1999, Cayer et al. 2007, Kirschner et al. 1999, Klausen et al. 1997, Matlova et al. 2005, mosses & Matlova et al. 2003, Reznikov and Leggo 1974, Wendt et al. 1980) dusts Potting soils (De Groote et al. 2006, Fujita et al. 2014, Fujita et al. 2013, Kaevska et al. 2011, Yajko et al. 1995) Environmental soils (Alfonso et al. 2004, Bannalikar and Verma 2007, Blacklock and Dawson 1979, Brooks et al. 1984b, Brooks et al. 1984c, Chilima et al. 2006, Costallat et al. 1977, Cvetnic et al. 1998, Dvorska et al. 2002, Eaton et al. 1995b, Fujita et al. 2014, Fujita et al. 2013, Gaynor et al. 1990, Jones and Jenkins 1965, Kaevska et al. 2011, Kamala et al. 1994a, Kamala et al. 1994b, Katila et al. 1995, Kleeberg and Nel 296 1969, Kleeberg and Nel 1973, Krizova et al. 2010, Matlova et al. 2003, Matthews et al. 1979, May et al. 1995, Parashar et al. 2004, Reznikov and Dawson 1980, Rhodes et al. 2013, Salgado et al. 2015, Schröder et al. 1992, Shitaye et al. 2008, Singer and Rodda 1965, Stanford and Paul 1973, Whittington et al. 2003, Wolinsky and Rynearson 1968, Yajko et al. 1995) Sand, wet stones & clay (Alqumber 2014) Sphagnum moss (Cooney et al. 1997, Schröder et al. 1992) Dust (Cvetnic et al. 1998, Dvorska et al. 2002, Ichiyama et al. 1988, Jin et al. 1984, Kaevska et al. 2011, Kamala et al. 1994a, Kamala et al. 1994c, Kleeberg and Nel 1969, Kleeberg and Nel 1973, Krizova et al. 2010, Matlova et al. 2003, Meissner and Anz 1977, Nel 1981, Reznikov and Dawson 1971, Reznikov et al. 1971b, Saitanu 1977, Torvinen et al. 2010, Tsukamura et al. 1974, Tsukamura et al. 1984, Viallier and Viallier 1973) Building materials / mold (Huttunen et al. 2000, Rautiala et al. 2004, Torvinen et al. 2006) from water damaged buildings 297 9.5. RHRW QMRA Table 9.20 Water uses reported by Currumbin (n = 45) and Brisbane (n = 76) rainwater survey participants (one survey entry per tank). Note that multiple uses were reported for each tank and percentages were calculated using the total number of study participants that responded to the question in each location (n = 76 and n = 44, for Currumbin and Brisbane, respectively). The survey was undertaken during March-September 2015 and reproduced from (Hamilton et al. 2016). Location Uses (%) Drinking Cooking Clothes Showering Pool Gardening Car Ornamental Toilet Fish Pet washing top-up washing water feature flushing tanks washing filling Currumbin (n 42 (100) 42 (100) 39 (93) 42 (100) 1 (2.4) 15 (36) 17 (41) 2 (4.8) 6 (14) 0 (0) 0 (0) = 42)* Brisbane (n = 15 (20) 8 (11) 19 (25) 9 (12) 25 (33) 69 (92) 33 (44) 0 (0) 25 (33) 2 (2.7) 2 (2.7) 75)* Total (n = 57 (49) 50 (43) 58 (50) 51 (44) 26 (22) 84 (72) 50 (43) 2 (1.7) 31 (27) 2 (1.7) 2 (1.7) 117)* *Sample numbers indicate number of participants who responded to the survey in each location and indicated a response to the rainwater usage question 298 9.6. A critical review of approaches for Legionella QMRA6 9.6.6. Abstract Legionella has been identified as the responsible agent for two-thirds of waterborne disease outbreaks in the United States from 2011-2012. The prevention of Legionella in engineered systems presents complex challenges for the drinking water industry due to its persistence, resistance to disinfection, and complex microbial ecology. Not all species of Legionella are of concern for human health, however, certain environmental conditions can cause human-virulent species such as L. pneumophila to proliferate, or modulate the distribution of virulence characteristics for relevant strains such that health risks are presented. Quantitative microbial risk assessment (QMRA) is a tool for integrating information on pathogen occurrence, infectivity, and exposure for guiding water quality management strategies. A standardized QMRA approach for Legionella has not been developed, and exposure models are highly varied based on scenario- and site-specific conditions. Detailed discussion of these varied mathematical approaches has been limited, but can aid in identifying research gaps for further QMRA development and public health prevention policies. A summary of 18 studies that utilize Legionella exposure models for sewage treatment plants, cooling towers, drinking water distribution systems, whirlpool spas, showering, and recreational water scenarios are discussed here. Ten of these studies conducted a full QMRA, and provided human infection estimates. The summarized models utilized Gaussian dispersion, volumetric estimation, occupational hygiene, and aerosol science approaches. Parameters, implications, and 6 This section has been published: Hamilton, K.A. and Haas, C.N. (2016) Critical review of mathematical approaches for quantitative microbial risk assessment (QMRA) of Legionella in engineered water systems: research gaps and a new framework. Environmental Science: Water Research & Technology 2(4), 599-613. 299 limitations of each of these mathematical approaches are discussed, and a QMRA framework to address the identified limitations is proposed. This framework provides a comprehensive overview of key steps within an idealized Legionella QMRA model from exposure to risk characterization, including biofilm impacts, aerosol generation, survival and transport of bacteria within size-resolved water droplets, and interaction with a human receptor. 9.6.7. Introduction Legionella is a genus of opportunistic bacteria with several species of significant public health importance known to occur in engineered water systems, ambient water environments and soils. Legionnaires’ disease or the milder form Pontiac fever result from inhalation or aspiration of aqueous aerosols or soil dusts containing Legionella bacteria(enHealth 2004). The most common cause of illness is L. pneumophila, although over 50 species have been identified (Diederen 2008). Infections are particularly problematic for susceptible populations such as those who are immunocompromised or immunosuppressed, the elderly, smokers, and other hospitalized individuals (Benami et al. 2016, Masters and Edwards 2015). Although rapid diagnosis and treatment methods have improved case-fatality rates since the first recognized outbreak in 1976, prevention of Legionnaires’ disease remains an important focus for water quality management (Fogarty et al. 2015). Legionella was identified as the responsible agent for 66% of reported waterborne disease outbreaks and 26% of waterborne illnesses in the United States from 2011-2012(Beer et al. 2015). The prevention of human-virulent Legionella spp. in engineered systems presents complex challenges for the drinking water industry. The presence of Legionella spp. does 300 not necessarily dictate health risk due to variability in the occurrence of human-virulent subsets and virulence characteristics within individual species. Legionella is relatively resistant to disinfection compared to other waterborne bacteria, and persists in piped distribution systems due to its ability to colonize biofilms(ABS 2010). Legionella can also proliferate and enhance its virulence when internalized into protozoan symbionts such as Acanthamoeba within biofilms. Portions of biofilms can become detached from surfaces and enter the bulk water, becoming dispersed in mists produced by various water fixtures (Reznikov et al. 1971a, Thomas et al. 2010). Elevated temperatures (20 - 45° C), increased water age, and the presence of algal deposits, sludge, or nutrients are associated with enhanced Legionella growth (September et al. 2004). In addition to controlling these factors, risk management challenges arise when choosing disinfectants that simultaneously mitigate multiple pathogen risks. For example, switching from free chlorine to monochloramines is more effective in preventing Legionella growth, but in some cases may favor colonization of the biofilm with Mycobacterium avium Complex (MAC), another potential opportunistic pathogen of public health concern that occurs in similar ecologic niches to Legionella(McSwiggan and Collins 1974, Pryor et al. 2004b, Van Ingen et al. 2009). Quantitative Microbial Risk Assessment is a framework that integrates information regarding pathogen occurrence, infectivity, and exposure for determining the health implications of microbial hazards. The QMRA is conducted using a process of hazard identification, exposure assessment, dose response assessment, and risk characterization(Haas et al. 2014). Conducting a QMRA for Legionella spp. can aid in identifying opportunities for prioritizing public health risk prevention and mitigation 301 strategies. Although guidance is available for conducting general QMRA frameworks(September et al. 2004), there is no standardized approach for developing a Legionella QMRA, and exposure models are highly varied based on scenario-specific conditions. These exposure models are further complicated by the need to address uncertainties in detection methodologies and identifying pathogenic subsets of Legionella bacteria, variability in environmental concentrations, viability and infectivity considerations, host immunity, and the fate and transport of bacteria in aerosols under varying environmental conditions (Whiley et al. 2014b). Detailed discussion of the assumptions and limitations of varied quantitative modeling approaches has been limited. This review summarizes published QMRA mathematical frameworks for Legionella (Section 3) and identifies strengths, limitations, and research gaps for further QMRA model development (Section 4). 9.6.8. Frameworks for Legionella risk assessment 9.6.8.1. General frameworks for microbial dispersion modeling & exposure analysis. The primary types of models for particle dispersion are simple box, Gaussian Plume, Lagrangian, and Eulerian (Holmes and Morawska 2006). Model selection typically depends on the desired spatial scale and complexity of the analysis (Van Leuken et al. 2015). Several reviews (Dungan 2010, Lighthart 1994, Van Leuken et al. 2015) have summarized fate and transport models for bioaerosols, which have relied heavily upon modified Gaussian Plume models. However, Gaussian Plume models are typically valid only for downwind exposures ranging from 100 m to 10,000 m (Lighthart 1994), 302 and other techniques are necessary for short-range exposures. QMRA models for wastewater, biosolids, and spread of dusts containing pathogens between farms have used Gaussian Plume models with various modifications (Brooks et al. 2012, Brooks et al. 2005b, Dowd et al. 2000, Galada et al. 2012, Jahne et al. 2015, Jahne et al. 2014, Ssematimba et al. 2012, Tanner et al. 2008, Teng et al. 2013, Viau et al. 2011), with several studies generating site specific, meteorological data-intensive estimates using United States Environmental Protection Agency (USEPA) AERMOD software (Dungan 2014, Jahne et al. 2015, Jahne et al. 2014) or Eulerian Computational Fluid Dynamics (CFD) (Blatny et al. 2011, Blatny et al. 2008, Fossum et al. 2012). Additionally, Lagrangian ballistic or random walk models have been used for modeling microbial dispersion where individual aerosol behavior is accounted for, taking into account phenomena such as deposition, evaporation, and bacterial decay (Falkinham et al. 2001a, Lighthart and Kim 1989). Gaussian Plume, empirical aerosol concentration estimates, and CFD approaches have been applied within the context of Legionella QMRA frameworks. A summary of 18 published Legionella exposure analysis and QMRA studies follows. Major study characteristics are listed in Table 9.21. 303 Table 9.21 Summary of Legionella exposure models reviewed Reference (Number) Model Exposure source Analytical method Biofilm Bacterial Aerosol size consideration typea used to determine consideration enrichment during generation or Legionella transport concentrationb Blatny et al. (2008) AS, CFD Biological treatment plant Culture, qPCRf x Fossum et al. (2012) AS, CFD Air scrubber NSe x Nygård et al. (2008) G Air scrubber PCRf x Nguyen et al. (2006) G Cooling tower NSe Rouil et al. (2004) G Cooling tower NSd Storey et al. (2004) V Drinking water distribution system NS x x Ahmed et al. (2010) V Showering, hosing qPCR x Armstrong and Haas (2007b) PC, NFFF Whirlpool spa, hot spring Culture x x Armstrong and Haas (2008) PC, NFFF Whirlpool spa, hot spring Culture x x Medema et al. (2004) PC Sewage treatment plant Culture, PCRc Schoen and Ashbolt (2011) PC Shower Culture x x de Man et al. (2014a) PC Recreational splash parks that use rainwater as a Culture, qPCR source Sales-Ortells and Medema (2015) PC Various urban water systems Culture, qPCR Sales-Ortells and Medema (2015) PC Stormwater plaza aerosols qPCR Schoen et al. (2014) NS Shower qPCR NS NS NS Schoen et al. (2014) NS Shower qPCR NS NS NS Azuma et al. (2013) PC Bathing in a residential bathroom Culture Bouwknegt et al. (2013) AS Whirlpool spa Culture x aV= volume estimation; PC= partitioning coefficient; NFFF= near-field far field; G=Gaussian plume or puff; CFD= computational fluid dynamics; AS= Aerosol science approach; NS= not specified; bRefers to the analytical method on which the concentration data is based where Culture= culture-based methods, PCR= polymerase chain reaction (binary, non-quantitative); qPCR= quantitative polymerase chain reaction; cLegionella were measured using both culture and PCR based methods and were not recovered from aerosols using the culture-based method, however only a qualitative risk evaluation was provided based on these data; dRouil et al. 2004 assumed that 1 µg /m3 aerosolized water was equivalent to 7 × 10-4 Legionella / m3 air (units unspecified), computations done in terms of µg / m3; eAuthors did not compute Legionella concentrations in plumes and modeled total aerosol concentrations as surrogate for potential risk; fEnvironmental investigation was conducted using culture and/or (q)PCR, however dispersion modeling was based on water aerosol transport only 304 9.6.8.2. Gaussian plume and puff models. Several studies (Nguyen et al. 2006, Nygård et al. 2008, Rouil et al. 2004) have used Gaussian models to develop exposure estimates in conjunction with Legionnaires’ disease outbreaks linked to cooling towers and are summarized elsewhere (Van Leuken et al. 2015). Although these studies did not conduct full QMRAs, a brief discussion of the Gaussian Plume model is presented here in equation 9.1, as this approach can be integrated with existing exposure approaches to calculate a deposited dose for QMRA, and is useful for comparison with the other approaches discussed below. A Gaussian Puff model further incorporates this equation with a Lagrangian trajectory model (Van Leuken et al. 2015). 2 푄 1 푦 1 푧−퐻 2 1 푧+퐻 2 −휆푥 퐶(푥, 푦, 푧) = 푒푥푝 [− ( ) ] [푒푥푝 {− ( ) } + 푒푥푝 {− ( ) }] 푒푥푝 [ ] 2휋푈휎푥휎푧 2 휎푦 2 휎푧 2 휎푧 푈 Equation 9.1 Where C= the concentration of bacteria in air at a downstream position with spatial coordinates x, y, and z; Q =emission rate [pathogens / s]; U= wind speed [m / s]; H= emission height [m]; and σy and σz are dispersion coefficients in the x and z directions [m]. These formulas can be modified for QMRA by multiplying by additional inhalation exposure parameters. Nguyen et al.Nguyen et al. (2006) compared dispersion estimates with geographic areas in France where 86 cases resided during an outbreak using the Atmospheric Dispersion Modeling System (ADMS) model. A qualitative correlation was observed 305 between cases and predicted concentrations. Rouil et al. Rouil et al. (2004) conducted a study to assess the source of a Legionella outbreak in Lens, France hypothesized to be linked to a cooling tower using ADMS. However, the authors concluded that the model only accounted for approximately 20% of the observed variance in infection rates. Nygård et al.Nygård et al. (2008) modeled a similar Norwegian outbreak with 56 cases using the Gaussian integrated puff (INPUFF) model and identified a biological treatment plant for further study. 9.6.8.3. Volumetric estimation approaches. One approach for Legionella exposure modeling assumes that a given activity produces a volume or range of volumes for aerosols in the respirable range. This volume is typically not broken down by aerosol size category, but rather aggregated by calculating the total volume of water per exposure event. Bacteria are then transferred uniformly to these aerosols and are transmitted to a receptor via inhalation as in equation 9.2. 퐷 = 퐶푤푎푡푒푟퐿푎푒푟퐼푅 ∗ 퐸푇 Equation 9.2 Where D= the dose of Legionella; Cwater = concentration of Legionella in bulk water [# / 3 L]; Laer = load of aqueous aerosols of respirable diameter in air [L / m ]; IR= inhalation rate [m3 / min] and ET= exposure time [min]. An early but comprehensive Legionella QMRA for drinking water distribution systems was developed by Storey et al.Storey et al. (2004). The approach considered 306 Legionella to be ubiquitously present and uniformly distributed between biofilms and bulk water, and proportionally distributed within aerosols. However, Legionella in the bulk water was not accounted for, and only Legionella introduction into water due to biofilm sloughing was considered. A triangular distribution was used for the concentration of Legionella in biofilms with minimum, mean, and maximum of 102, 103, and 106, respectively (microbiological units not specified). Biofilm detachment and disinfection were modelled using experimental trials from stainless steel and polyvinyl chloride coupons in annular reactors (Storey and Ashbolt 2003). A shower peak concentration of 0.1- 10 µm size aerosols was used to compute an average volume inhaled over the course of a 10-minute inhaled exposure of 414 ± 258 µg / m3 or 57.5 ± 35.8 µL. All Legionella inhaled were assumed to be deposited, with 365 day per year exposure. A conservative “maximum risk” curve of r = 1 was used as a dose response model for L. pneumophila was not developed until 2007 (Armstrong and Haas 2007a). The maximum risk curve concept signifies that each microbial exposure will result in an infection. The simulations indicated that if 100 Legionella / cm2 biofilm is present with no disinfection during a 10% sloughing event, the probability of infection per event is 0.18 and the annual risk is 1 for inhalation of 50 µL water. It was concluded that the maximum density of Legionella in biofilms during a 10% sloughing event and inhalation of 50 µL of water that met USEPA guidelines was 0.05 Legionella / cm2. Treatment of water to 80° C for 10 min reduced risks to below a 1:10,000 benchmark (du Moulin et al. 1985). Ahmed et al. Ahmed et al. (2010) assessed Legionella risk from inhalation of aerosols due to daily showering or twice weekly hosing with domestic roof-harvested 307 rainwater in Southeast Queensland, Australia. The concentration of L. pneumophila was measured in 214 roof-harvested rainwater samples collected from 82 tanks using quantitative polymerase chain reaction (qPCR), assuming that one gene copy was equivalent to 1 viable, infective Legionella cell as there is a single copy of the L. pneumophila mip gene per genome. The concentrations ranged from 60 -170 gene copies (gc) / L (5.6% rainwater tanks positive). A survey conducted determined that 4.06% and 1.01% of households used rainwater for hosing and showering, respectively. It was assumed that all households with rainwater tanks used water for hosing, but only the percentage designated as potable would use the water for showering. To obtain the volume transmitted via aerosols for showering, the aerosol size distribution next to a shower and hose were identified (Keating and McKone 1993, O'Toole et al. 2008b, Zhou et al. 2007) and corrected for lung deposition efficiencies (Schlesinger 1985) considering only aerosols in the 0.3 - 6.0 µm diameter range. Using an inhalation rate of 20 L / min for an adult undergoing “light activity” for a 7-minute hot shower, the volume inhaled was calculated to range from 0.02 µL to 0.84 µL during showering (0.84 µL worst case used), 0.008-0.04 µL during high-pressure hosing, and 0.09-0.5 µL during hosing with trigger nozzle settings (a worst case of 0.5 µL was used). Using the infection dose response model for Legionella (r = 0.06) (Armstrong and Haas 2007a), the number of infections per 10,000 exposed people per event ranged from 3.0 × 10-2 - 8.6 × 10-2 for showering and from 1.8 × 10-2 - 5.1 × 10-2 for hosing. The annual infection risk per 10,000 Southeast Queensland residents ranged from 2.6 × 10-3 - 7.3 × 10-3 for showering and 2.1 × 10-3 - 5.8 × 10-3 for hosing. 308 9.6.8.4. Occupational & industrial hygiene approaches Several concepts used in Legionella QMRA have been developed from the occupational and industrial hygiene literature for chemical exposure, including the use of near field-far field (NFFF) models and partitioning coefficients (PC) (Feazel et al. 2009, Huang et al. 2010). NFFF models assume well-mixed concentrations in a room, but address spatial variability in exposure by dividing the exposure zone into two areas, the near-field zone containing the source, and the far-field zone box model that exchanges air with the near-field zone. NFFF models account for ventilation (dilution), but not particle settling or viability decay(Armstrong and Haas 2007b). The partitioning coefficient (also known as an “emission factor” or “efficiency”) concept is similar to the volumetric approach, except that this is a direct measurement of the ratio between bacteria measured in water and air (equation 9.3). 퐶 푃퐶 = 푎𝑖푟 Equation 9.3 퐶푤푎푡푒푟 Where PC= partitioning coefficient; Cair= concentration of Legionella in air 3 [microorganisms / m ]; and Cwater= concentration of Legionella in water [microorganisms / L]. Partitioning coefficients are summarized in Hines et al.Hines et al. (2014b) and range from 1.3 ×10-6 L / m3 (toilets) to 8.8 × 10-1 L / m3 (cool mist humidifiers) for various inhalation exposures. A general model using the partitioning coefficient or exposure factor is stated in equation 9.4 (Hines et al. 2014b). 309 퐷 = 퐶푤푎푡푒푟 ∗ 푃퐶 ∗ 퐼푅 ∗ 퐸푇 Equation 9.4 Where D= exposure dose (colony forming units [CFU] or count / L); PC= partitioning -3 -1 3 coefficient [CFU m / CFU L , also written as L / m ]; Cwater= water concentration [CFU or count / L]; IR= inhalation rate [m3 / min]; and ET = exposure time [min]. Medema et al.(Medema et al. 2004) conducted a Legionella risk assessment for sewage treatment workers exposed to aerosols using data gathered from five treatment plants. Seven wastewater samples were tested and although 7/7 were positive using PCR (up to 105 / mL), samples were negative using culture-based methods (detection limits < 20 - < 200 / mL). Legionella spp. and L. pneumophila concentrations in air samples ranged from 0.56 - 56 gc / m3 in 3/5 treatment plants, but negative using culture-based methods. The value of the partitioning coefficient was not reported for Legionella, but was stated to be similar to the efficiency for heterotrophic plate counts of 10-10 - 10-4.5 presented at the same site. If computed from the PCR data, the partitioning coefficient would range from 5.6 × 10-6 - 5.6 × 10-4 L / m3. Using survey data on the activity patterns of sewage workers, exposure doses were estimated to be 23 Legionella / day, and 5% of the workers were exposed to 25 Legionella with a maximum exposure of 37 Legionella. The annual risk was not computed, however, the risk potential was considered low when compared to concentrations observed in the literature at other outbreak-associated sites. Schoen and Ashbolt (2011) used a reverse QMRA model to identify the target concentration of Legionella in premise plumbing biofilms necessary to deliver a target dose of 1-100 CFU associated with infection in the lower respiratory tract. Exposure was 310 considered during a 15 minute shower operating at a flow rate of 6 L / min and inhalation rate of 0.72 - 1.5 m3 / h using the exposure model in equation 9.5. A partitioning coefficient of 1 × 10-6- 1 × 10-5 CFU m-3 / CFU L-1 was applied. In addition, a fraction of bacteria that partitioned proportionally into aerosols of size 1 - 5 µm (0.75 - 1), 5 - 6 µm (0.09 - 0) and 6 - 10 µm (0.14) diameter was considered and corrected for the proportion of aerosols in each of these ranges that deposited at the alveoli (0.2 - 0.54, 0.1 - 0.65, and 0.01 - 0.1, respectively). 퐿 퐷퐷 퐷푤푎푡푒푟 = 1 2 Equation 9.5 푃퐶∗퐼푅∗퐸푇 ∑ 퐹𝑖 퐹𝑖 퐿 Where 퐷푤푎푡푒푟= the concentration of Legionella in water necessary to achieve the target deposited dose [CFU / L]; DD= target deposited dose [CFU]; PC= partitioning -3 -1 3 1 coefficient [CFU m / CFU L ]; IR= inhalation rate [m / h]; 퐹푖 = the fraction of total 2 aerosolized organisms in aerosols of size i; 퐹푖 = the fraction of aerosols of size range i that deposit at the alveoli. This approach assumed that the biofilm was the sole source of Legionella in the bulk liquid, and that exposure would occur as biofilm material was detached, released through the trophozite or cyst form of protozoa located in the biofilm, or released within vacuoles or vesicles derived from those protozoan hosts. The behavior of Legionella under these conditions was modeled in biofilms prior to a detachment event by considering the fraction and intensity at which protozoan trophozoites become infected with Legionella (0.01-1 and 10-1000 CFU / host, respectively), the total biofilm surface area (40 - 4000 cm2), the fraction of area sloughed off from the biofilm during a 311 detachment event (100% and all of respirable size), and the sloughing rate (520 - 1560 g / cm2) (equations 9.6-9.9) 퐿 퐷∗퐹푅∗퐸푇 퐷푆퐴 = Equation 9.6 퐹푏𝑖표푓𝑖푙푚∗푆퐴 퐷∗퐹푅 퐷퐿 = Equation 9.7 푀 퐷푅∗푆퐴 퐿 ℎ표푠푡 퐷푆퐴 퐷푆퐴 = 3 퐶 Equation 9.8 퐹ℎo푠푡∗퐿ℎ표푠푡 퐿 ℎ표푠푡 퐷푀 퐷푀 = 3 퐶 Equation 9.9 퐹ℎ표푠푡∗퐿ℎ표푠푡 퐿 2 퐿 Where 퐷푆퐴 = the number of organisms per unit surface area of biofilm [CFU / cm ], 퐷푀= 2 ℎ표푠푡 the number of organisms per unit mass of biofilm [CFU / g ]; 퐷푆퐴 = the number of hosts 2 ℎ표푠푡 (amoebae) per unit surface area of the biofilm [host / cm ]; 퐷푀 = the number of hosts 2 (amoebae) per unit mass of biofilm [host / g ]; FR= shower flow rate [L / h]; Fbiofilm= the fraction of biofilm surface area that is sloughed off and of respirable size; SA= total 2 2 3 biofilm surface area [cm ]; DR= biofilm detachment rate [g / cm ]; 퐹ℎ표푠푡= trophozite 퐶 infection fraction; and 퐿ℎ표푠푡= Legionella infection intensity [CFU / host]. Predicted critical densities of Legionella ranged from 3.5 × 101 - 3.5 × 103 CFU / m3 in air, 3.5 ×106 - 3.5 × 108 in water, 7.8 × 105 - 7.8 × 108 CFU /cm2 (by surface area) or 5 × 102 – 5 × 105 CFU / g (by mass) in the biofilm. Within the biofilm, predicted protozoan host densities ranged from 3.1 ×1 04 - 7.8 × 107 hosts / cm2 and 2.0 × 101 - 5.0 312 × 104 hosts / g. The median target deposited dose of 10 CFU was based on the median infectious dose (LD50) of 12 CFU from the exponential dose response model for guinea pig infection with r= 0.06, but did not make use of the model function expressly (Norton and LeChevallier 2000b). A sensitivity analysis demonstrated that the changes in the partitioning coefficient and target deposited dose contributed most to uncertainty in the critical density of Legionella in water. In a risk assessment for Legionella-containing rainwater used at recreational splash parks, de Man et al. de Man et al. (2014a) extracted values from the literature for the concentration of L. pneumophila with an average concentration of of 1200 CFU / L (Gamma distribution (r = 0.045, λ = 26,000), where CFU= colony forming units. An exposure volume of 0.394 µL / min (95% confidence interval 0.0446- 1.27 µL / min) for children and 0.489 µL / min (95% confidence interval 0.0494-1.55µL / min) for adults was used with a mean duration of 3.5 minutes at an interactive water fountain. The exposure volume was derived by multiplying the inhalation rate and a partitioning coefficient. Although the study specifies a “volume of inhalable water spray (VIWS)”, this is distinct from the volumetric method described previously as the VIWS. The VIWS (average 10.8, 95% confidence interval 1.76- 36.3 µL / m3) was estimated using maximum likelihood and beta regression to calculate a ratio of measured concentrations of inhalable endotoxin in air and water near splash parks, which varied from 7.2 to 19 endotoxin units / m3 air and 9 to 2799 endotoxin units / mL water (endotoxin units cancel; note the partitioning coefficient units are frequently reported in both formats, for example CFU m-3/ CFU L-1, or L / m3). The correlation between endotoxin units in water and air was significant (R2 = 0.645). Inhalation rates of 1.01 × 10-2- 4.36 ×10-2 m3 / min 313 for children and 1.03 × 10-2 - 7.77 × 10-2 m3 / min for adults were used to obtain the exposure volume [µL / min]. The exposure duration was based on an observational study at two splash parks in urban centers, which observed 257 children and 347 adults within 2 m of a water spray. An exponential dose response model with an infection endpoint was used with r = 0.06 (Armstrong and Haas 2007a). The mean risk per 3.5 min exposure duration was 9.3 × 10-5 (95% confidence interval 0 - 2.4 × 10-4) for children and 1.1 × 10- 4 (0 - 82.8 × 10-4) for adults. The authors note that the duration could be much longer in a recreational water park environment, up to 0.5 h or possibly up to 2 h. Therefore, for a 2 h exposure, the risk would be 2.8 × 10-3. The results of the scenario analysis indicated that the volume of inhalable water spray was the most important input parameter for determining Legionella risk. Sales-Ortells and Medema (2014)(Sales-Ortells and Medema 2014) performed a screening-level microbial risk assessment of urban water locations including a river, canals, a lake, and a water playground located in the Netherlands. Twenty water bodies or features were identified by the Amersterdam water utility for potential for human exposure. Study sites were identified based on human exposure potential to contaminated waters, obtained using expert judgement. These sites were ranked, and the top 13 chosen for further study. The selected features were of categories of: 1) local storage of stormwater runoff in wide infiltration trenches (“wadi’s”); 2) urban green/blue areas with temporary storage of stormwater from separated sewers in ponds and ditches; 3) urban water recreation areas such as water playgrounds, water fountains, and local surface water used for recreation; and 4) water on the street during rain events. Scenarios for which Legionella risks were calculated were aerosol exposures due to: 1) rowing on a 314 river; 2) rowing on a river influenced by combined sewer overflow (CSO) contamination; 3) rowing on a lake; 4) rowing on a lake influenced by separate sewer overflow (SSO) contamination; 5) public water taps; 6) ornamental fountains; 7) drinking water playgrounds; and 8) fresh water playgrounds influenced by runoff from traffic roads. The aerosolization ratio (partitioning coefficient) reported by de Man et al.(de Man et al. 2014b) (log transformed parameters mean ± standard deviation -8.07, 0.3) was used to estimate the concentration of L. pneumophila in air for each scenario, derived from decorative fountains. Note that these values are slightly different to those reported in de Man et al.(de Man et al. 2014a) due to a different statistical fitting procedure to the original data. Due to the sparse nature of Legionella concentrations for these types of water systems, concentrations of Legionella from the literature were pooled and fit with gamma distributions for a lake that was impacted by a wastewater discharge rich in L. pneumophila, and one scenario excluding those data. A 1-log reduction in bacteria based on E. coli data was used to describe removal in a sedimentation pond, and a triangular distribution (0.1, 1, 2) was used to describe pathogen log reductions due to dilution and natural processes for canals in a green area and park. A 10-fold dilution for wastewater was considered for CSO-impacted waterways. Inhalation rates were calculated for each exposure activity and age group from the USEPA 2011 Exposure Factors Handbook but specific values were not stated (USEPA 2011a). Inhalation rates were corrected using a lung deposition efficiency in the lower respiratory tract of 12.7%. Exposure durations and frequencies were activity specific, where rowing had the longest duration (1 – 4 hours per event and up to 108 events per 315 year for river rowing). Infection (r = 0.06) and illness (r = 1.7 × 10-4) dose response models were used. The highest legionellosis risks per event were calculated from playing in pluvial floodwater impacted by a CSO (1 × 10-2), rowing on the river (1.4 × 10-5) or lake (1.4 × 10-5), and playing at the surface water playground (3.4 × 10-6). All annual risk median for Legionella were below the national annual incidence of legionellosis (0.002% average 2009-2010). These risks were higher than those observed for roof-harvested rainwater systems in Australia, and the concentrations used were higher; however the authors point out variations in aerosol estimation and exposure methods that would limit comparisons between the two studies. According to the sensitivity analysis, the variability in legionellosis risk was higher than that for gastrointestinal pathogens; this was driven nearly entirely by the variability in Legionella concentrations. The studies used to fit L. pneumophila concentration distributions were derived from a mixture of culture-based (Australian Government Department of Health 2016a), qPCR (Phin et al. 2014), culture and qPCR (de Man et al. 2014a) or unspecified methods (Albrechtsen 2002) from various water systems. Lesser but moderate spearman rank correlations were shown for the impact of exposure frequencies on legionellosis risk. The other inputs showed weak correlations with legionellosis risk. Sales-Ortells and Medema (2015) similarly modeled microbial risks from Legionella due to inhalation of stormwater aerosols in a stormwater plaza in Bellamyplein, The Netherlands. The plaza collects street runoff and diverts the first flush into a sewer system before storing the remainder in an open area. L. pneumophila was measured in the stormwater plaza using qPCR and was quantified in 2 of 10 pond samples in concentrations ranging from 1.5 × 102 to 1.1 × 103 gc / L. A gamma 316 distribution was fitted to the concentration data and a beta distribution to the recovery efficiency (average 13.2- 25.8%) which was used to correct the concentration data. The same aerosolization ratio approach derived from de Man et al.(de Man et al. 2014b) (log transformed parameters mean ± standard deviation -8.07, 0.3) was used to estimate the concentration of L. pneumophila in air. An inhalation rate for children was used (1.36 L / min) for an exposure time of 21 ± 5 minutes and a lung deposition efficiency of 12.7%. Infection (r = 0.06) and illness (r = 1.7 × 10-4) dose response models were used. As the stormwater plaza will only fill when there is sufficient rainfall, a 10-year rainfall record was used to describe a negative binomial distribution for the frequency of exposure events. From 2004 - 2013, 14.6 rainfall events were equal to the fill volume and 38.3 were higher than this volume. The calculated dose was 1.1 × 10-5 gc (95% 5.21 × 10-5 gc) and the risk per person per event was 1.2 × 10-9 (95% 5.2 × 10-9). The measured pathogen concentration and aerosolization ratio had the highest impact on the risk estimates. Schoen et al. (2014) modeled the annual probability of illness from inhalation of treated rainwater while showering for L. pneumophila in order to compare decentralized community water services to conventional centralized services. qPCR concentration data for rainwater was obtained from published studies (Ahmed et al. 2014b, Ahmed et al. 2010) . The exposure model was not specified, although based on previous work (Schoen et al. 2011). An exponential dose response model for illness was used (r = 1.07E-4) (Armstrong and Haas 2008), resulting in annual probabilities of illness that were generally between 10-6 to 10-9. Similarly, Schoen and Garland (2015) conducted a review to identify treatment reduction targets for onsite water reuse. Legionella was used as a 317 reference hazard for showering in untreated rainwater, however the predicted annual probability of infection was less than 10-3 per person-year. Armstrong and Haas (2007b) and Armstrong and Haas (2008) Armstrong and Haas (2008) address aspects of the same two exposure models, for a whirlpool spa and a natural hot spring thermal spa. The natural hot spring thermal spa analysis is based on data gathered from two hot spring outbreaks in the Miyazaki and Shizuoka Prefectures in Japan. A mean partitioning coefficient derived from endotoxin data from swimming pools without water features of 2.3 × 10-5 (90% range 1.6 × 10-5- 3.1 × 10-5) was used to estimate occupational exposure to Legionella-containing mists. Reported concentrations from Miyazaki and Shizuoka sampling campaigns were 1.5 × 107 CFU / L and 7 × 105 CFU / L, respectively, and air sampling was not conducted. Therefore, the mean estimated air concentrations using the partitioning coefficient were 360 CFU / m3 (95% CI 240 - 470) for Miyazaki and 17 CFU / m3 (95% CI 12 - 23) for Shizuoka. No decay or loss of viability for Legionella was considered in the short time between emission and inhalation for a 15-minute exposure duration. A uniform distribution of inhalation rates ranging from 0.6 - 1.5 m3 / h was used, consistent with a light to moderate activity level. An estimated 50% of aerosols were considered to be in the respirable range and deposited at the alveoli, resulting in a calculated mean deposited dose of 47 CFU (95% confidence interval 23 -78 CFU) for Miyazaki and 2.3 CFU (95% confidence interval 1.1-3.7 CFU) for Shizuoka. Using the exponential dose response model for infection (r = 0.06), the worker subclinical infection risks for Miyazaki and Shizuoka were 7.5 × 10-1 - 9.9 × 10-1 and 6.5 × 10-2 - 2.2 × 10-1, respectively. Using an exponential model (r = 8.7 × 10-5) with an animal death endpoint, the clinical severity Legionnaires’ disease risk was 2.0 × 10-3- 318 7.3 × 10-3 and 9.8 × 10-5 - 3.5 × 10-4 for Miyazaki and Shizuoka, respectively. The sensitivity analysis indicated that the partitioning coefficient and breathing rate contributed most to dose estimate variability. The whirlpool spa exposure model is based on data gathered at a large outbreak that occurred at the 1999 Netherlands West Frisian Floral Show, with 133 confirmed and 55 probable cases of LD(Tortoli et al. 2004). A whirlpool spa injects air into spa water, forming an emission plume. A near-field far-field steady state model with estimates for aerosol generation, water composition, and building ventilation parameters was used to calculate estimates for near field (2 – 15 m) and far field (>15 m) occupational exposures to this plume. The equations for near field concentration, far field concentration, and air exchange rate between both zones are shown in equation 9.10-9.12. 퐺 퐺 퐶 = + Equation 9.10 푁퐹,푆푆 푄 훽 퐺 퐶 = Equation 9.11 푓퐹,푆푆 푄 푆 훽 = Equation 9.12 2퐹푆퐴 3 Where CNF,SS= the near field concentration at steady-state [CFU / m ]; CFF,SS= the far field concentration at steady state [CFU / m3]; G= microbial emission rate of the contaminant [CFU / min]; Q= the room supply/exhaust ventilation rate (m3 / min); β= the air exchange rate between the near and far field zones (m3 / min); FSA= the free surface 319 are of the assumed near field zone (m2) s= the average random, non-directional air velocity between the near and far field zones (m / min). Surface area was calculated assuming a cylindrical near field zone. Equation 9.13 was used to obtain the microbial generation rate (G): 1 퐺 = 퐶 퐷 퐶 퐸 퐸퐹 Equation 9.13 휌 푎푒푟 푎푒푟 푤푎푡푒푟 푎푖푟 Where G= the generation rate [CFU / min]; Caer= the concentration of aerosols in the 3 respirable range (2 - 5 µm) above the whirlpool; ρ= density of water [g / m ]; Daer= 3 aerosol load [µL / m ]; Cwater= concentration in bulk water [CFU / mL]; Eair= air emission rate 0.38 m3 / min; and EF= enrichment factor [dimensionless]. The pumping capacity of a typical whirlpool spa was estimated at 380 - 760 L / min, which given a 50/50 spa air to water ratio would result in an emission plume of 0.19 - 0.38 m3 / min. Data from an aerosol mass distribution above a whirlpool spa in the 1 to 7 µm aerosol diameter range indicated an aerosol load of 5 × 10-3 mL / m3 (Baron and Willeke 1986). Enrichment, or increased concentration of an aqueous aerosol relative to the bulk water concentration of gram negative bacteria, has been reported (Blanchard and Syzdek 1982, Chang et al. 2003) . Therefore, two scenarios of zero enrichment and a 10-fold enrichment were derived from an estimate for surface foam in a cooling tower basin of 1000-fold enrichment (Dawson 1971). A value of 10 or 1 % of this value was used because the mixing in a whirlpool is more likely to create a homogeneous distribution than for cooling towers. It was assumed that the Legionella in water aerosol equals the concentration in the spa’s bulk water, no decay or settling of Legionella occurred within 320 the plume, and the whirlpool bulk water would not be significantly depleted of Legionella by aerosol emissions. With no enrichment, the aerosol generation rate was 2 - 7 CFU / min, and for 10-fold it was 20 - 70 CFU / min. Exposures of 30 h ± 21.8 h were used, considering the total number of hours for workers during the outbreak period at the flower show. The inhalation rate was 0.6 - 1.5 m3 / h. Fifty-percent aerosol retention was assumed based on a guinea pig model(Armstrong and Haas 2007a), resulting in a deposited dose of 0.25 - 34.4 CFU for workers within 15 m, and 0.24 - 18.5 for workers greater than 15 m away from the source. Subclinical and clinical severity dose response models were with r = 0.06 and r = 1.07 × 10-4, respectively. Mean subclinical infection risks ranged from 5.8 × 10-2 (no enrichment) to 3.9 × 10-1 (enrichment) for distances less than 15 m and 4.0 × 10-2 - 3.2 × 10-1 for distances greater than 15 m. Mean clinical severity risks ranged from 8.9 × 10-5 - 8.9 × 10-4 and 6.0 × 10-5 – 6 × 10-4 for distances below and above 15 m, respectively. The results of a sensitivity analysis with rank order correlation coefficients indicated that inhalation rate, exposure time, and the microbial aerosol generation rate were the most important for determining the infection risk. Azuma et al. (2013) used information from a 2007 outbreak in Adachi, Japan to determine the appropriateness of the Japanese water quality guideline for Legionella of 100 CFU/ L. Using the same approach as Armstrong and Haas (Armstrong and Haas 2007b) for hot springs, a partitioning coefficient was used for assessing the risk of Legionella from bathing in residential bathrooms (Uniform distribution 2700 -13,000 CFU / L). However, a Poisson distributed exposure duration (λ = 21 min) was used. A mean of 7850 CFU / L Legionella was measured in a hot spring water circulation system 321 in a condo building in Adachi, Tokyo. The mean calculated concentration in air was 0.18 CFU / L (95% confidence interval 0.07 - 0.32), resulting in a mean deposited dose of 0.033 (95% confidence interval 0.010 - 0.069). The mean risk of infection per bath was 2.0 × 10-3 (95% confidence interval 0.6 - 4.1 × 10-3) while the mortality risk was 2.8 ×10- 6 (95% confidence interval 0.9 × 10-6 – 6 ×10-6). Mean annual risks were 5.1 × 10-1(2.0 × 10-1 - 7.8 × 10-1) and 1 × 10-3 (0.3 × 10-3 × 10-3 - 2.2 × 10-3) for infection and mortality dose response endpoints, respectively. The infection and mortality risks of the water quality guideline value 100 CFU / L were approximately 1 × 10-2 and 1 × 10-5. These findings supported a water quality guideline of 1 CFU / L. 9.6.8.5. Aerosol science approaches. A more detailed treatment of processes occurring during Legionella transport are possible using aerosol science approaches that consider the physical forces acting on individual water droplets. This is useful for situations such as whirlpools and cooling towers, where dynamic aerosol behavior occurs through mechanisms of settling, evaporation, condensation, coagulation, interaction among different components of particles, and formation of secondary aerosols (Nygård et al. 2008). Few studies have considered this approach, as it typically requires more data inputs and is more computationally intensive than the other approaches discussed. Risks from L. pneumophila from whirlpool use were quantified by Bouwknegt et al. (2013). An injected air stream produces jet bubbles that ascend rectilinearly to the water surface and intercept Legionella in the bulk water at certain efficiencies. As the jet bubbles meet the water surface, they produce larger jet drops as well as small, inhalable 322 film drops. The rise of air bubbles through a water column spanning the depth of the whirlpool was modeled. During bubble rise, the number of Legionella scavenged from the bulk water was estimated using interception efficiency rates. Assuming that equal portions of jet drops and film drops are produced, half of the Legionella were estimated to become entrained in film droplets and the concentration of Legionella in both the whirlpool water and air above the whirlpool were assumed to be homogenous. 1×106 identical droplets of 0.71 cm diameter were simulated using a previously published aerosol size distribution (Baron and Willeke 1986), and an air concentration was computed in a cylindrical air column above the whirlpool defined by the dimensions of the pool and height of the ceiling. These factors were combined with breathing rates to achieve an inhaled dose. A combined form of the multiple exposure model equations presented in the paper is provided in equation 9.14. 푅푉 퐼퐹 6 퐶 푉 퐸 푓푎 2 푠 푠 ∑10 푤푎푡푒푟 푏 푎 Equation 9.14 퐷 = 푗=1 [1 + 37 0.85] 푉푎𝑖푟 퐴 1+( ) 푅푒 Where D= inhaled dose of Legionella [CFU]; RVs= respiratory volume [L] with s indicating gender-specific volumes; IFs= inhalation frequency [n per minute] with s indicating gender-specific frequencies; Vair= volume of the air column above the whirlpool [mL]; Cwater= concentration of Legionella in water; Vb= volume of water 2 traversed by bubbles moving rectilinearly according to Vb = πr h where r is the radius of the bubble and h is the depth of the whirlpool; Ea= the probability of attachment given interception, assumed here to be unity; f = the fraction of Legionella contained in film droplets; a = the radius of Legionella spp.; A = the radius of uniform bubbles; Re = 323 휌휈퐿 Reynold’s number, calculated according to 푅푒 = where ρ = the density of water, ν= µ the mean velocity of rising bubbles, L= the bubble diameter, and µ= the dynamic viscosity of water. Film drops from a circular whirlpool of 3 m diameter, 70 cm depth, and water level 1 m from the ground were simulated using the air injection volume and 50% air- water mixture previously reported by Armstrong and Haas (2007b) with airflow of 360 - 780 L / min. The slowest velocity in the path of a bubble noted by Aybers and Tapucu Norton et al. (2004) of 23.7 cm / s was used, and assumed to be independent of water temperature. The calculated efficiency of bubbles encountering Legionella in the water column was 8.4 × 10-4. The calculated generation rate was 1 - 2×106 air bubbles / min and aerosol depletion was assumed to be equivalent to Legionella growth, so that the Legionella in the whirlpool would not be depleted. Various conditions for concentration (1 - 1×106 CFU / L), inhalation rate (9 L / min for males and 6.7 L / min for females), and exposure duration (1 min, 15 min, or 2 h) were investigated. Equation 9.14 resulted in approximately 0.01% of the total Legionella in the whirlpool inhaled by a receptor, however this fraction was concentration- dependent. For starting water concentrations of 1000 CFU / L, the bacterial generation rate was 4.5×104 CFU / min (95% confidence interval 4.4 × 104 - 4.7 × 104) and the average inhaled dose ranged from 58 CFU (males) to 43 CFU (females). Using the exponential model for infection (r = 0.06), infection risks for a 10 CFU/ L starting water concentration ranged from 0.002 (males, 1 min exposure) to 0.24 (males, 2 h exposure). Exposure to water containing ≥ 1000 CFU / L nearly always caused infection in both males and females. A scenario analysis assessed the impacts of turbulence (changing the 324 Reynolds number and modifying the type of motion from rectilinear to helical), the fraction of intercepted Legionella that becomes airborne, and a ventilation scenario (homogeneous versus nonhomogeneous mixing in the air column). Results indicated that the fraction of intercepted L. pneumophila on film drops and turbulence of water in the whirlpool affecting the bubble path have the most influence on infection risks. The authors indicated support for a limit of 100 L. pneumophila CFU / L for bathing water. With regard to other aerosol science approaches, CFD models have been applied to calculating Legionella exposures but did not calculate infection risks and are summarized elsewhere (Van Leuken et al. 2015). Three studies (Blatny et al. 2011, Blatny et al. 2008, Fossum et al. 2012) investigated an air scrubber at a biological treatment plant in Norway identified as a probable source of airborne Legionella in an outbreak by Gaussian modeling efforts (Nygård et al. 2008). Blatny et al. (2008) and Fossum et al. (2012) used a steady state Reynolds Averaged Navier-Stokes (RANS) approach and noted that the mean plume concentration of Legionella 200 m downwind was approximately 2% of the source strength. It was also observed in a related study that bacteria were principally contained by small (< 4 µm) or large (> 16µm) aerosol size fractions (Blatny et al. 2011). 9.6.9. Limitations of current models and research needs for QMRA development Eighteen studies are presented here that provide exposure models for Legionella in engineered water systems using Gaussian dispersion, volumetric estimation, occupational hygiene, and aerosol science approaches. Ten of these models conducted a full QMRA, and provided a full exposure model description and human infection estimates. A 325 comparison of risk characterization parameters for studies that conducted full QMRAs is shown in table 9.22. This comparison demonstrates that ranges of inhalation rates reported across studies are typically narrow (0.081 - 4.662 m3 / h), and two principal dose response model functions have been used for infection and clinical severity infection or death. Beyond variability accounted for in the dose response function, variation in host susceptibility has not been considered within the risk characterization of these studies. However, differences in risk characterization parameters are small compared to the differences in parameterization across exposure models for similar exposure scenarios. Among studies that conducted sensitivity analyses, the primary factors identified as contributors to infection risk estimates are parameters from the exposure model, particularly with regards to the volume of water that becomes inhaled and number of bacteria that are transferred to air (microbial generation rates, water volumes inhaled, and partitioning coefficient)(Armstrong and Haas 2007b, Buse et al. 2012, Schoen and Ashbolt 2011). However, inhalation rate and concentrations of Legionella in water sources were also identified as important factors. Note that the factors that appear in a sensitivity analysis are a function of the type of assessment; for example for the reverse QMRA conducted by Schoen and Ashbolt, the concentration of Legionella would not appear in the sensitivity analysis as this was the outcome variable, however Legionella concentration was deemed an important variable by Sales-Ortells (2015). Limitations specific to each type of exposure model are discussed in the following sections. 326 Table 9.22 Risk characterization parameters for reviewed models that conducted full QMRAs for which a full model description was available Reference Exposed Inhalation Rate Exposure Duration (min) No. exposure eventsb Dose response parameters Parameters of populationa (m3 / h) importance identified by sensitivity analysis Storey et al. Gen 0.83 10 365 Maximum risk curve r = 1 ND (2004) Ahmed et al. Gen 1.2 7 1; Annual risk with showering daily, Infection r = 0.06 ND (2010) hosing twice weekly Armstrong and Occ 0.6-1.5 Whirlpool: 1800 ± 1260 1 Subclinical infection r = 0.06; Inhalation rate, work Haas (2007b) severity Legionnaires’ disease hours, microbial aerosol or mortality r = 1.7×10-4 generation rate Armstrong and Occ 0.6-1.5 Hot spring: 15 1 Subclinical infection r = 0.06; Breathing rate, PC Haas (2008) severity Legionnaires’ disease or mortality r = 1.7×10-4 Schoen and Gen 0.72 (best 15 1 Target deposited dose of 10 PC, target deposited Ashbolt (2011) estimate) - 1.5 CFU based on ID50 of 12 dose (high estimate) CFU from infection dose response model (r = 0.06) de Man et al. Gen Children: 0.606 3.5 1 Infection r = 0.06 Volume of inhalable (2014a) - 3.744d Adults: water spray, inhalation 0.618 -4.662 rate Sales-Ortells and Gen NSf e,gRowing: T (60, 120, 240); e,gRowing in river: SU (1, 108); rowing Infection r = 0.06; severity Concentration of L. Medema (2015) Playing in rainwater reservoirs/ in lake: NB (5.1, 12); swimming in Legionnaires’ disease or pneumophila in water street runoff: N (21, 5); Playing in fresh water: NB (8, 1.3); swimming in mortality r = 1.7×10-4 fresh water playground: LN (4.1, swimming pool water: NB (24, 1); 0.80); playing in drinking water wading/splashing in flooded streets: NB playground: LN (4.2, 0.55); (8, 2); wading/splashing in water walking on flooded street/car playgrounds/ornamental fountains: B splashing: T (0.017, 0.083, 0.17); (12, 0.2); walking the dog, walking walking dog: T (15, 30, 60); public close to public taps: B (12, 0.2) water taps: point estimate 1 Sales-Ortells and Gen Normal (0.081, 21±5 1 Subclinical infection r = 0.06; Pathogen concentration, Medema (2015) 0.0036)e severity Legionnaires’ disease aerosolization ratio or mortality r = 1.7×10-4 Azuma et al. Gen 0.6-1.5 Poisson (21, 4.55)c 1 per day. Risks simulated per event, Subclinical infection r = 0.06; ND (2013) per 2 months, and per year severity Legionnaires’ disease or mortality r = 1.7×10-4 Bouwknegt et al. Gen 0.54 (male), Simulated for 1, 15, 120 1 Infection r = 0.06 Fraction of intercepted (2013) 0.402 (female) bacteria on film drops; turbulence of water a Occ=Occupational; Gen= general population bFor exposure events where 1 event was indicated, risk estimates were provided on a “per event” basis; ND= Not done c Poisson parameters shown are mean, standard deviation dInhalation rate ranges for light activity (low end of range) to high intensity activity (high end of range) eParameters of the normal or lognormal distribution are (µ, σ) fNot specified gT= triangular distribution (min, mode, max); SU= step uniform, parameters (min, max); B= binomial, parameters (N, P); NB= negative binomial, parameters (µ, k) 327 9.6.9.1. Gaussian plume and puff models. The Gaussian Plume model assumes that the mean concentration of airborne particles are normally distributed about the axis downwind from a contaminant point source, and the concentration gradient diminishes as downwind distance and distance away from the plume centerline increases (Lighthart 1994). Although the Gaussian Plume is one of the most straightforward models that can be used for long-range dispersion, it is applicable generally only for downwind distances from a point source 100 – 10,000 m, and for conditions when meteorological and source conditions are stable for periods longer than the distance of the receptor from the source divided by the mean wind speed (Lighthart 1994). The empirical constants σy and σz represent standard deviations of the plume concentration in the horizontal and vertical directions, respectively. These constants are based on averages during experiments performed under various atmospheric stability and relatively short time duration conditions (many authors have derived methods for determining these constants, but the most widely used are those by Turner (Morales- Pinzón et al. 2015), based on experimental data from Pasquill (Chidamba 2015)), and are therefore not universally applicable to all contaminants and environmental conditions despite their widespread use for highly varied scenarios. The impact of ground-level topography is generally not addressed by the Gaussian Plume model, and more computationally intensive approaches (such as CFD) are required for these types of applications. In addition to the effect of dispersion coefficients, estimates are highly sensitive to the effective stack height (H). For point sources (most Legionella sources are likely to be point sources) such as cooling towers, stack height would be higher than the actual 328 physical stack height used for simulation in most models due to the effects of plume rise. Plume rise occurs because the plume is hotter than the surrounding air, and rises buoyantly as it exits the stack with a vertical velocity (Thomson et al. 2013c). In order to calculate plume rise, it is necessary to obtain specific information regarding the stack height exit velocity, stack diameter, and temperature of exiting water vapor; therefore this has not typically been included in previous generalized models of bioaerosol dispersion. However, increasing the value used for stack height will increase the downwind dispersion of a contaminant, and should be included for site-specific calculations where feasible. Finally, the decay rates for Legionella and other microorganism transport for use in combination with the Gaussian Plume model have been derived by numerous authors, typically under controlled laboratory conditions and as functions of temperature and relative humidity, but rarely taking into account the effects of light (UV) inactivation, limiting their application for realistic transport simulations (Berendt 1980, 1981, Dennis and Lee 1988, Hambleton et al. 1983, Katz and Hammel 1987, Lighthart 1973, 1989, 1994, Lighthart and Frisch 1976). These factors become more important as travel time from the source to a receptor increases (Lighthart 1994). Together, these limitations provide many challenges for comparing modeled microorganism dispersion estimates with epidemiological information on the spatial location of Legionnaires’ disease cases in relation to a point source. Van Leuken et al. (Van Leuken et al. 2015) re-analyzed predicted concentrations and attack rates for qualitative data reported by Nguyen et al. (Nguyen et al. 2006) with a linear regression function, and noted that although a 329 correlation between Legionella dispersion during an outbreak was positive it was not significant (p ≈ 0.12). 9.6.9.2. Volumetric estimation approaches. The volumetric estimation approach assumes that systems or activities produce a given volume of water spray that can be inhaled by a human receptor. Volumetric approaches are useful, especially for screening level assessments, as they are straightforward to apply and can be performed with a low burden of data inputs. The approach is versatile in that it does not need to make assumptions about the behavior of microorganisms during transport, but can easily incorporate such data, such as the differential partitioning of bacteria into the liquid water, or into aqueous aerosols of various sizes. This can be considered a strength of the approach, especially in establishing QMRA frameworks for easy incorporation of additional data once it becomes available. However, all of the studies that used the volumetric approach considered a homogeneous distribution of Legionella in aerosols initially as well as during transport between the water fixture/system and receptor. 9.6.9.3. Occupational & industrial hygiene approaches Most exposure studies reviewed modeled either the transport of water droplets or of bacteria, but did not combine these approaches. The partitioning coefficient was the most common approach reported for use in QMRA (8 /10 of the reviewed studies that conducted a full QMRA). The partitioning coefficient is a ratio of bacterial concentrations in air and water, and similarly to volumetric approaches, is a straightforward and relatively versatile approach for considering the airborne fraction of 330 Legionella bacteria, especially for screening-level assessments. However, the ratio of microorganisms in water and air has been shown to change with bubble rise distance and bulk water concentration (Jensen et al. 2002, Torvinen et al. 2004a, Viau and Peccia 2009). The partitioning coefficient also does not consider the behavior of aerosol droplets. A large range of generally low partitioning coefficient values were reported across studies (ranging from 10-10- 10-4.5 CFU m-3 / CFU L-1). However, the partitioning coefficient typically does not specify the distance at which the ratio of bacteria in air and water were measured, making it difficult to interpret what fraction of airborne bacteria are within the respirable range at varying distances from a bacterial emission source. This makes it more difficult to combine the partitioning coefficient approach with other methods compared to the volumetric approach, where a volume associated with an initial aerosol size distribution can be used and applied in conjunction with other assumptions to develop a reasonable estimate of Legionella doses. Schoen and AshboltSchoen and Ashbolt (2011) assumed a fractional partitioning of Legionella into aerosols of various sizes which could be used in combination with multiple exposure model approaches, however, the original submitted manuscript that was cited has not been published. The partitioning coefficient approach implicitly assumes that the ratio of bacteria in the respirable range remains constant over space and time. The use of partitioning coefficients or volumetric calculations alone does not account for aerosol size considerations in most cases. The enrichment of individual aerosol droplets is also not considered. Although partitioning coefficient values are low, indicating that only a fraction of bacteria in the bulk water become aerosolized, the increased concentration due to enrichment that occurs in small aerosol droplets in the initial aerosol size distribution is 331 likely to play an important role in disease transmission, especially as smaller particles are less likely to settle during plume transport and could deliver higher doses to receptors than predicted. Blatny et al.(Blatny et al. 2011) emphasized the importance of considering the size distribution as Legionella were likely to be found in mostly small (< 4 µm) or large (>16 µm) droplets (Van Leuken et al. 2015). 9.6.9.4. Aerosol science approaches Determining the aerosol size distribution and downwind proportion of Legionella- containing aerosols in the respirable range remains a substantial challenge for QMRA models. Aerosol science methodology provides the most detailed and mechanistic approach for modeling Legionella dispersion from engineered water systems. However, these approaches are likely to present challenges in the form of data needs and computational requirements. It is challenging to model the evolution of aerosol size distributions over time due to co-occurring and interrelated dynamic rate physical phenomena of settling, evaporation, condensation, coalescence, and secondary aerosol formation due to bubble burst and film collapse (Hinds 1999, Lighthart et al. 1991). The fraction of aerosols in the respirable range is therefore not likely to remain constant over time. Nygård et al. (2008), Bouwknegt et al. (2013), and Armstrong and Haas (2007b) made simplifying assumptions by using a single size index aerosol within the respirable range. However, bacteria are likely to survive better in an aggregated form containing biofilms and/or protozoan hosts, which are more likely to persist in larger droplet nuclei over time (Lighthart and Kim 1989, Lighthart and Shaffer 1997). These larger droplets could potentially reach a respirable size by the time they arrive at a receptor distance due to evaporation. Additional robustness conferred to Legionella bacteria during transport by 332 biofilm elements has not been studied, and the assumption of an index aerosol droplet of small size may not be sufficient (Blatny et al. 2011). Furthermore, the impact of environmental conditions on these factors remains a significant research gap. Droplets containing Legionella bacteria should be constrained to a minimum size of the diameter of Legionella (1 – 2 µm)(Armstrong and Haas 2007b). 9.6.9.5. Research gaps Due to the ubiquitous nature of Legionella in engineered water systems, continual exposure is likely. Three studies (Sales-Ortells and Medema 2014, 2015, Storey et al. 2004) considered infection risks over more than one exposure period, and reported risks using an exposure frequency by annualizing risk estimates. However, Armstrong and Haas (2007b) considered total individual worker exposure time over the course of several days. The authors concluded that determining human time and activity patterns can reduce uncertainty in risk estimates, especially when combined with better ventilation information for indoor exposures. The nature of most Legionella QMRA and environmental investigations have been in conjunction with outbreaks, therefore the applicability of these risk estimates to normal operating conditions may be limited (Bouwknegt et al. 2013). In addition to variability in actual Legionella concentrations under various circumstances, variability in measured Legionella concentrations is likely to be impacted by the analytical detection method used. More than half (6 of 10) of the full QMRA studies reviewed based their assessments of Legionella concentrations on determinations using culture-based data. It has been shown that molecular assays (PCR or qPCR) are more sensitive than culture- based assays as they are capable of quantifying both live and dead cells, and are not 333 limited by non-target microorganism overgrowth and viable but non-culturable (VBNC) issues that hinder culture-based methods (Lee et al. 2016). Furthermore, cell starvation, exposure to heat or disinfectants, and passage through amoebic cells can render Legionella cells in a VBNC state, but nevertheless alive and capable of resuscitation (Amin and Han 2011, Dobrowsky et al. 2015a, Dobrowsky et al. 2015b, Lee and Visscher 1992, Reyneke et al. 2016). Accordingly, a review of culture and qPCR approaches by Whiley and Taylor(Whiley and Taylor 2014) demonstrated that 26 of 28 reviewed studies detected Legionella at a higher rate using qPCR compared to culture, and on a sample per sample basis, samples analyzed concurrently by qPCR and culture were approximately 50% more likely to return a positive result by qPCR than by culture. One of the studies reviewed here quantified Legionella using both culture and qPCR methods (Medema et al.), finding 0.56- 56 gc/ L using qPCR and an absence of the organism using culture. These findings support that the use of qPCR data has the potential to yield higher calculated risks in a QMRA than if concentration data generated using culture-based methods are used. Although it is common in QMRA studies to pool concentration estimates from the literature where microbiological measurements of pathogens such as Legionella in a specific type of water source are limited, this approach could potentially introduce additional sources of variability into concentration estimates. Sales-Ortells et al.and de Man et al. (de Man et al. (2014a), Sales-Ortells and Medema 2014) calculated Legionella data from studies that reported using both culture- and qPCR- methods. Although it is not clear if culture and qPCR data were pooled within a single concentration distribution, Sales-Ortells et al. (2014) noted that the Legionella concentration was the dominant factor in determining legionnellosis risks during a 334 sensitivity analysis. It is difficult to determine the extent to which pooling sources of Legionella data generated using different microbiological methods could have an impact on risk estimates, however, it appears that this practice could potentially introduce issues in interpretation of sensitivity analyses. Reporting detailed information regarding the derivation of concentration distributions used in a QMRA would therefore be useful for interpretation, especially as sensitivity analyses are used to prioritize the collection of further datasets. In the case of pathogen concentration data, this could imply significant additional costs for collection of such data. A correction factor parameter could be applied within the context of a QMRA to estimate the viable and/or infectious fraction for quantification of health risks in the case of using a non-culture-based dataset. However, a singular factor is not likely to be universally applicable across all conditions and therefore warrants further development and review, especially during aerosol generation and transport. Low or highly variable recovery rates are frequently reported, especially for culture- based methods (Brooks et al. 2005a). One reviewed study (Sales-Ortells and Medema 2015) corrected measured concentrations with a beta distribution for recovery efficiency. The use of uncorrected reported concentrations reported in most of the reviewed studies may underestimate risks and correcting for recovery efficiency is an important practice. However, directly correcting for recovery efficiency can also overestimate concentrations and introduce bias (Gale 2005). A Bayesian framework was recommended by Schmidt et al. (Martinson and Thomas 2009) to address these issues for protozoans, and similar approaches could be applied for Legionella (Gale 2005). Legionella viability, infectivity, virulence state, and survival in transport as a function of environmental conditions 335 (relative humidity, ultraviolet light, temperature, and protection by organic debris or biofilms) are additional areas for further study to improve QMRA models. In addition to limitations associated with concentration estimates, the dose response models used in all of the reported studies was derived from culture data. Therefore, there may be some limitations in applying this model in conjunction with a qPCR dataset. This discrepancy could lead to an overestimation of qPCR risks consistent with the limitations mentioned above; i.e. as a dose-response function for culture-based data is likely to have a lower median infectious dose (ID50) than a qPCR-based dataset for the same microorganism. Therefore, a dose response model using a culture- and qPCR- method comparison is recommended as a research gap that needs to be addressed. Some limitations could presumably be overcome by recreating the laboratory conditions under which the culture-based dose response models were generated and quantifying the inoculums using both culture and qPCR methods. In the QMRAs reported here, the ability to directly compare risk estimates across studies is limited due to the variety of parameterizations across exposure models and varied starting concentrations of Legionella. However, the importance of the behavior of Legionella during the aerosolization process appears to be a recurrent theme across QMRAs. An ideal Legionella QMRA model would dynamically track the evolution in size distribution, and simultaneously address the fate of bacteria contained within aqueous droplets. The fate of bacteria would also be determined as a function of environmental conditions between a bacterial source and receptor. A framework for such a model that considers the limitations discussed in this review is proposed in figure 1. Many elements of the reviewed studies can be utilized and integrated to form the basis of 336 this approach. Although more rigorous approaches are needed for treatment of aerosol size, these approaches are likely to require more data inputs and be more computationally intensive. However, a better integrative understanding of mathematical frameworks for Legionella modeling can lead to more rigorous risk prevention strategies and water management approaches. Furthermore, the QMRAs reviewed here highlight the importance of emerging sources of Legionella exposure that warrant further investigation, including alternative and/or decentralized water sources such as rainwater harvesting(Ahmed et al. 2010) and stormwater management infrastructure systems(Sales- Ortells and Medema 2014, 2015, Schoen and Garland 2015). 337 Recovery- corrected Bacterial enrichment for concentration of Legionella from biofilms Aerosol generation rate systems in which bubbles Legionella in water and enter bulk water (scenario specific) or condensed mists are biofilms generated Bacterial survival, Aerosol transport, virulence, viability during Size-resolved bacterial Size distribution of dispersion, evaporation transport as a function of partitioning in aerosol generated aerosol and deposition environmental conditions Size resolved deposition Risk characterization & Inhalation Dose response model of aerosol at the alveoli sensitivity analysis Figure 9.9 Proposed comprehensive framework for Legionella quantitative microbial risk assessment 338 9.7. Legionella reclaimed water QMRA 9.7.6. Cooling towers Stack Ht (m) Endpointa Culture qPCR 10 Inf D 100 Inf D Distance from source (m) aInfection (Inf) or Death (D) animal dose response model endpoints; death corresponds to clinical severity infection. Note that in some cases the 5th percentile did not converge due to a low fraction of samples positive. Figure 9.10 Annual infection risks (10y) for Legionella pneumophila in residential populations due to cooling tower aerosol exposure using drinking water at varying downwind distances for wind speed = 7 m / s and relative humidity = 65%. The median (solid black line) and 95% confidence interval (dotted lines) are shown. 339 Stack Ht (m) Endpointa Culture qPCR 10 Inf D 100 Inf D Distance from source (m) aInfection (Inf) or Death (D) animal dose response model endpoints; death corresponds to clinical severity infection. Note that in some cases the 5th percentile did not converge due to a low fraction of samples positive. Figure 9.11 Annual infection risks (10y) for Legionella pneumophila in occupational populations due to cooling tower aerosol exposure using drinking water at varying downwind distances for wind speed = 7 m / s and relative humidity = 65%. The median (solid black line) and 95% confidence interval (dotted lines) are shown. 340 Location Culture- residential Culture-occupational 7 27 30 31 32 33 341 Figure 9.12 Annual infection risks (10y) for Legionella pneumophila due to cooling tower aerosol (0.001- 0.005% efficiency) exposure at varying downwind distances for wind speed = 7 m / s, relative humidity = 65%, stack height=10 m, infection dose response endpoint. The median (solid black line) and 95% confidence interval (dotted lines) are shown. Note that in some cases the 5th percentile did not converge due to a low fraction of samples positive. 342 Figure 9.13 Impact of cooling tower drift efficiency (left: 0.001-0.005%; right: 0.01-0.1%) on annual infection risks (10y) for Legionella pneumophila in reclaimed water for residential populations due to cooling tower aerosol exposure at varying downwind distances for culture data, wind speed = 7 m / s, relative humidity = 65%, and stack height=10 m. The median (solid black line) and 95% confidence interval (dotted lines) are shown. 343 9.7.7. Sprinklers Endpointa Culture qPCR Inf D aInfection (Inf) or Death (D) animal dose response model endpoints; death corresponds to clinical severity infection. Note that in some cases the 5th percentile did not converge due to a low fraction of samples positive. Figure 9.14 Annual infection risks (10y) for Legionella pneumophila in residential populations due to sprinkler aerosol exposure using drinking water at varying downwind distances for wind speed = 7 m / s and relative humidity = 65%. The median (solid black line) and 95% confidence interval (dotted lines) are shown. 344 Endpointa Culture qPCR Inf D aInfection (Inf) or Death (D) animal dose response model endpoints; death corresponds to clinical severity infection. Note that in some cases the 5th percentile did not converge due to a low fraction of samples positive. Figure 9.15 Annual infection risks (10y) for Legionella pneumophila in occupational populations due to sprinkler aerosol exposure using drinking water at varying downwind distances for wind speed = 7 m / s and relative humidity = 65%. The median (solid black line) and 5th, 95th percentiles (dotted lines) are shown. 345 Location Culture- residential Culture-occupational 7 27 30 31 32 33 346 Figure 9.16 Annual infection risks (10y) for Legionella pneumophila due to sprinkler aerosol exposure at varying downwind distances for wind speed = 7 m / s, relative humidity = 65%, infection dose response endpoint. The median (solid black line) and 5th, 95th percentiles (dotted lines) are shown. aNote that in some cases the 5th percentile did not converge due to a low fraction of samples positive. 347 10. References ISO 11731-2:2006 Dansk Standard Water quality-Detection and enumeration of Legionella-Part 2: Direct membrane filtration method for waters with low bacterial counts. Abbasi, T. and Abbasi, S. (2011) Sources of Pollution in Rooftop Rainwater Harvesting Systems and Their Control. Critical Reviews in Environmental Science and Technology 41(23), 2097- 2167. Abrahams, E. and Harland, R. (1968) Studies with the purified protein derivative of human and avian tuberculins in South Queensland. Tubercle 49(2), 192-209. ABS (2010) Environmental issues: water use and conservation, Australian Bureau of Statistics, Canberra, Australia. Adams, A., Garbett, M., Rees, H. and Lewis, B. (1978) Bacterial aerosols from cooling towers. Journal (Water Pollution Control Federation), 2362-2369. Adams, A., Garbett, M., Rees, H. and Lewis, B. (1980) Bacterial aerosols produced from a cooling tower using wastewater effluent as makeup water. Journal (Water Pollution Control Federation), 498-501. Addiss, D., Davis, J., LaVenture, M., Wand, P., Hutchinson, M. and McKinney, R. (1989) Community acquired Legionnaires’ disease associated with a cooling tower: evidence for longer- distance transport of Legionella pneumophila. American journal of epidemiology 130(3), 557- 568. Adékambi, T., Salah, S.B., Khlif, M., Raoult, D. and Drancourt, M. (2006) Survival of environmental mycobacteria in Acanthamoeba polyphaga. Applied and Environmental Microbiology 72(9), 5974-5981. Adjemian, J., Olivier, K.N., Seitz, A.E., Falkinham, J.O., Holland, S.M. and Prevots, D.R. (2012a) Spatial Clusters of Nontuberculous Mycobacterial Lung Disease in the United States. American Journal of Respiratory and Critical Care Medicine 186(6), 553-558. Adjemian, J., Olivier, K.N., Seitz, A.E., Holland, S.M. and Prevots, D.R. (2012b) Prevalence of nontuberculous mycobacterial lung disease in U.S. Medicare beneficiaries. Am J Respir Crit Care Med 185. 348 Adrados, B., Julián, E., Codony, F., Torrents, E., Luquin, M. and Morató, J. (2011) Prevalence and concentration of non-tuberculous mycobacteria in cooling towers by means of quantitative PCR: a prospective study. Current microbiology 62(1), 313-319. Agdestein, A., Johansen, T.B., Kolbjørnsen, Ø., Jørgensen, A., Djønne, B. and Olsen, I. (2012) A comparative study of Mycobacterium avium subsp. avium and Mycobacterium avium subsp. hominissuis in experimentally infected pigs. BMC Veterinary Research 8(1), 11. Ahammed, M.M. and Meera, V. (2010) Metal oxide/hydroxide-coated dual-media filter for simultaneous removal of bacteria and heavy metals from natural waters. Journal of hazardous materials 181(1), 788-793. Ahmed, W., Brandes, H., Gyawali, P., Sidhu, J. and Toze, S. (2014a) Opportunistic Pathogens in Roof-Captured Rainwater Samples, Determined Using Quantitative PCR. Water Res 53, 361-369. Ahmed, W., Brandes, H., Gyawali, P., Sidhu, J.P.S. and Toze, S. (2014b) Opportunistic pathogens in roof-captured rainwater samples, determined using quantitative PCR. Water Res 53, 361-369. Ahmed, W., Gardner, T. and Toze, S. (2011a) Microbiological Quality of Roof-Harvested Rainwater and Health Risks: A Review. Journal of Environmental Quality 40(1), 13-21. Ahmed, W., Hamilton, K., Vieritz, A., Powell, D., Goonetilleke, A., Hamilton, M. and Gardner, T. (2016) Microbial risk from source-separated urine used as liquid fertilizer in sub-tropical Australia. Microbial Risk Analysis. Ahmed, W., Harwood, V., Gyawali, P., Sidhu, J. and Toze, S. (2015) Comparison of Concentration Methods for Quantitative Detection of Sewage-Associated Viral Markers in Environmental Waters. Appl Environ Microbiol 81(6), 2042-2049. Ahmed, W., Hodgers, L., Masters, N., Sidhu, J.P., Katouli, M. and Toze, S. (2011b) Occurrence of intestinal and extraintestinal virulence genes in Escherichia coli isolates from rainwater tanks in Southeast Queensland, Australia. Appl Environ Microbiol 77(20), 7394-7400. Ahmed, W., Huygens, F., Goonetilleke, A. and Gardner, T. (2008) Real-time PCR detection of pathogenic microorganisms in roof-harvested rainwater in Southeast Queensland, Australia. Appl Environ Microbiol 74(17), 5490-5496. Ahmed, W., Richardson, K., Sidhu, J., Jagals, P. and Toze, S. (2014c) Inactivation of faecal indicator bacteria in a roof‐captured rainwater system under ambient meteorological conditions. J Appl Microbiol 116(1), 199-207. 349 Ahmed, W., Vieritz, A., Goonetilleke, A. and Gardner, T. (2010) Health risk from the use of roof- harvested rainwater in Southeast Queensland, Australia, as potable or nonpotable water, determined using quantitative microbial risk assessment. Appl Environ Microbiol 76(22), 7382- 7391. Ahmed, W.V., A; Gardner, T; Goonetilleke, A (2009) Microbial risks from rainwater tanks in South East Queensland. water, 60-66. Aksamit, T.R. (2003) Hot tub lung: infection, inflammation, or both? Seminars in respiratory infections 18(1), 33-39. Al Sulami, A., Al Taee, A. and Wida'a, Q. (2012) Isolation and identification of Mycobacterium avium complex and other nontuberculosis mycobacteria from drinking-water in Basra governorate, Iraq. Alaimo, K., Packnett, E., Miles, R.A. and Kruger, D.J. (2008) Fruit and vegetable intake among urban community gardeners. J Nutr Educ Behav 40(2), 94-101. Albrechtsen, H. (2002) Microbiological investigations of rainwater and graywater collected for toilet flushing. Water Science and Technology 46(6-7), 311-316. Alfonso, R., Romero, R., Diaz, A., Calderon, M., Urdaneta, G., Arce, J., Patarroyo, M. and Patarroyo, M. (2004) Isolation and identification of mycobacteria in New World primates maintained in captivity. Veterinary Microbiology 98(3), 285-295. Alli, O.T., Olusoga, O.D., Adedokun, S. and Ogundare, O. (2011) Isolation of Legionella pneumophila from surface and ground waters in Osogbo, Nigeria. African Journal of Microbiology Research 5(18), 2779-2785. Alqumber, M.A. (2014) Prevalence of mycobacteria in water reservoirs of Albaha, Saudi Arabia. Saudi Medical Journal 35(5), 466-471. Alsup, S.E., Ebbs, S.D., Battaglia, L.L. and Retzlaff, W.A. (2011) Heavy metals in leachate from simulated green roof systems. Ecological Engineering 37(11), 1709-1717. Altare, F., Durandy, A., Lammas, D., Emile, J.-F., Lamhamedi, S., Le Deist, F., Drysdale, P., Jouanguy, E., Döffinger, R. and Bernaudin, F. (1998) Impairment of mycobacterial immunity in human interleukin-12 receptor deficiency. Science 280(5368), 1432-1435. Álvarez, J., García, I.G., Aranaz, A., Bezos, J., Romero, B., de Juan, L., Mateos, A., Gómez- Mampaso, E. and Domínguez, L. (2008) Genetic diversity of Mycobacterium avium isolates 350 recovered from clinical samples and from the environment: molecular characterization for diagnostic purposes. Journal of clinical microbiology 46(4), 1246-1251. Amemura-Maekawa, J., Kura, F., Helbig, J.H., Chang, B., Kaneko, A., Watanabe, Y., Isobe, J., Nukina, M., Nakajima, H. and Kawano, K. (2010) Characterization of Legionella pneumophila isolates from patients in Japan according to serogroups, monoclonal antibody subgroups and sequence types. Journal of medical microbiology 59(6), 653-659. Amin, M. and Han, M. (2011) Microbial quality variation within a rainwater storage tank and the effects of first flush in Rainwater Harvesting (RWH) System. Australian Journal of Basic and Applied Sciences 2011(5.9), 1804-1813pp. Angenent, L.T., Kelley, S.T., St Amand, A., Pace, N.R. and Hernandez, M.T. (2005) Molecular identification of potential pathogens in water and air of a hospital therapy pool. Proceedings of the National Academy of Sciences of the United States of America 102(13), 4860-4865. Angrill, J., Agustí, C. and Torres, A. (2001) Bronchiectasis. Current Opinion in Infectious Diseases 14(2), 193-197. Aoki, C., Memon, M. and Mabuchi, H. (2005) Water and wastewater reuse: an environmentally sound approach for sustainable urban water management. United Nations Environmental Program. Arasteh, K., Cordes, C., Ewers, M., Simon, V., Dietz, E., Futh, U., Brockmeyer, N. and L'age, M. (2000) HIV-related nontuberculous mycobacterial infection: incidence, survival analysis and associated risk factors. European journal of medical research 5(10), 424-430. Arbeit, R.D., Slutsky, A., Barber, T.W., Maslow, J.N., Niemczyk, S., Falkinham, J.O., O'Connor, G.T. and von Reyn, C.F. (1993) Genetic diversity among strains of Mycobacterium avium causing monoclonal and polyclonal bacteremia in patients with AIDS. Journal of Infectious Diseases 167(6), 1384-1390. Armstrong, T. and Haas, C.N. (2007a) A Quantitative Microbial Risk Assessment Model for Legionnaires' Disease: Animal Model Selection and Dose‐Response Modeling. Risk Analysis 27(6), 1581-1596. Armstrong, T.W. and Haas, C.N. (2007b) Quantitative microbial risk assessment model for Legionnaires' disease: assessment of human exposures for selected spa outbreaks. J Occup Environ Hyg 4(8), 634-646. Armstrong, T.W. and Haas, C.N. (2008) Legionnaires' disease: evaluation of a quantitative microbial risk assessment model. J Water Health 6(2), 149-166. 351 Aronson, T., Holtzman, A., Glover, N., Boian, M., Froman, S., Berlin, O., Hill, H. and Stelma, G. (1999) Comparison of large restriction fragments of Mycobacterium avium isolates recovered from AIDS and non-AIDS patients with those of isolates from potable water. Journal of clinical microbiology 37(4), 1008-1012. Asano, T., Franklin Burton, Harold Leverenz, Ryujiro Tsuchihashi, and George Tchobanoglous (2007) Water Reuse: Issues, Technologies, and Applications, McGraw Hill, New York. Ashbolt, N.J., Schoen, M.E., Soller, J.A. and Roser, D.J. (2010) Predicting pathogen risks to aid beach management: the real value of quantitative microbial risk assessment (QMRA). Water Research 44(16), 4692-4703. ASHRAE (2004) HVAC Systems and Equipment, Atlanta, GA. Athanasiadis, K., Helmreich, B. and Horn, H. (2007) On-site infiltration of a copper roof runoff: Role of clinoptilolite as an artificial barrier material. Water Research 41(15), 3251-3258. Athanasiadis, K., Helmreich, B. and Wilderer, P.A. (2004) Elimination of zinc from roof runoff through geotextile and clinoptilolite filters. Acta hydrochimica et hydrobiologica 32(6), 419-428. Australian Government Bureau of Meteorology (2016) Climate Data Online. Australian Government Department of Health (2016a) Australian national notifiable diseases and case definitions. Australian Government Department of Health (2016b) Notifications of a selected disease by month and year. AWWA (1999) Committee Report: Emerging Pathogens--bacteria, American Water Works Association. Azuma, K., Uchiyama, I. and Okumura, J. (2013) Assessing the risk of Legionnaires’ disease: The inhalation exposure model and the estimated risk in residential bathrooms. Regulatory Toxicology and Pharmacology 65(1), 1-6. Babrak, L., Danelishvili, L., Rose, S.J. and Bermudez, L.E. (2015) Microaggregate-associated protein involved in invasion of epithelial cells by Mycobacterium avium subsp hominissuis. Virulence 6(7), 694-703. 352 Baird, J.M., Summers, R. and Plummer, R. (2013) Cisterns and safe drinking water in Canada. Canadian Water Resources Journal (ahead-of-print), 1-14. Balian, S., Pinheiro, S., Guerra, J., Morais, Z., Ferreira, F., Oliveira, E. and Neto, J.F. (2003) INFECÇÃO EXPERIMENTAL EM HAMSTER (MESOCRICETUS AURATUS) COM MYCOBACTERIUM AVIUM. ESTUDO BACTERIOLÓGICO E HISTOLÓGICO. Arq. Inst. Biol., São Paulo 70(4), 397-405. Band, J., Ward, J., Fraser, D., Peterson, N., Silcox, V., Good, R., Ostroy, P. and Kennedy, J. (1982) Peritonitis due to a Mycobacterium chelonei-like organism associated with intermittent chronic peritoneal dialysis. Journal of Infectious Diseases 145(1), 9-17. Bannalikar, A.S. and Verma, R. (2007) PCR-RFLP analysis (PRA) of non tuberculous mycobacteria (NTM) isolated from animal specimens and environment using hsp65 and rpoB genes. Indian Journal of Animal Sciences 77(12), 1211-1218. Bardana, E., McClatchy, J., Farr, R. and Minden, P. (1973) Universal occurrence of antibodies to tubercle bacilli in sera from non-tuberculous and tuberculous individuals. Clinical and Experimental Immunology 13(1), 65. Barker, A.F. (2002) Bronchiectasis. New England Journal of Medicine 346(18), 1383-1393. Barker, J. and Jones, M. (2005) The potential spread of infection caused by aerosol contamination of surfaces after flushing a domestic toilet. J Appl Microbiol 99(2), 339-347. Barker, S.F., O'Toole, J., Sinclair, M.I., Leder, K., Malawaraarachchi, M. and Hamilton, A.J. (2013) A probabilistic model of norovirus disease burden associated with greywater irrigation of home-produced lettuce in Melbourne, Australia. Water Res. Baron, P.A. and Willeke, K. (1986) Respirable droplets from whirlpools: measurements of size distribution and estimation of disease potential. Environmental research 39(1), 8-18. Barrabeig, I., Rovira, A., Garcia, M., Oliva, J., Vilamala, A., Ferrer, M., Sabrià, M. and Domínguez, A. (2010) Outbreak of Legionnaires' disease associated with a supermarket mist machine. Epidemiol Infect 138(12), 1823-1828. Bartram, J., Chartier, Y., Lee, J., Pond, K. and Surman-Lee, S. (2007) Legionella and prevention of legionellosis. World Health Organization. 353 Bauer, J., Andersen, Å.B., Askgaard, D., Giese, S.B. and Larsen, B. (1999) Typing of clinical Mycobacterium avium complex strains cultured during a 2-year period in Denmark by using IS1245. Journal of clinical microbiology 37(3), 600-605. Beer, K.D., Gargano, J.W., Roberts, V.A., Hill, V.R., Garrison, L.E., Kutty, P.K., Hilborn, E.D., Wade, T.J., Fullerton, K.E. and Yoder, J.S. (2015) Surveillance for waterborne disease outbreaks associated with drinking water—United States, 2011–2012. MMWR Morb Mortal Wkly Rep 64, 842-848. Behets, J., Declerck, P., Delaedt, Y., Creemers, B. and Ollevier, F. (2007) Development and evaluation of a Taqman duplex real-time PCR quantification method for reliable enumeration of< i> Legionella pneumophila in water samples. Journal of microbiological methods 68(1), 137- 144. Benami, M., Busgang, A., Gillor, O. and Gross, A. (2016) Quantification and risks associated with bacterial aerosols near domestic greywater-treatment systems. Science of The Total Environment 562, 344-352. Benin, A.L., Benson, R.F. and Besser, R.E. (2002) Trends in legionnaires disease, 1980–1998: declining mortality and new patterns of diagnosis. Clinical infectious diseases 35(9), 1039-1046. Benjamini, Y. and Hochberg, Y. (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the royal statistical society. Series B (Methodological), 289-300. Bennett, S.N., Peterson, D.E., Johnson, D.R., Hall, W.N., Robinson-Dunn, B. and Dietrich, S. (1994) Bronchoscopy-associated Mycobacterium xenopi pseudoinfections. American Journal of Respiratory and Critical Care Medicine 150(1), 245-250. Berendt, R.F. (1980) Survival of Legionella pneumophila in aerosols: effect of relative humidity. Journal of Infectious Diseases 141(5), 689-689. Berendt, R.F. (1981) Influence of blue-green algae (cyanobacteria) on survival of Legionella pneumophila in aerosols. Infection and immunity 32(2), 690-692. Bermudez, L. (1994) Immunobiology of Mycobacterium avium infection. European Journal of Clinical Microbiology and Infectious Diseases 13(11), 1000-1006. Bermudez, L., Wu, M., Enkel, H. and Young, L.S. (1989) Naturally occurring antibodies against Mycobacterium avium complex. Annals of Clinical & Laboratory Science 19(6), 435-443. 354 Bermudez, L.E., Petrofsky, M., Kolonoski, P. and Young, L.S. (1992) An Animal Model of Mycobacterium avium Complex Disseminated Infection after Colonization of the Intestinal Tract. The Journal of Infectious Diseases 165(1), 75-79. Berry, D., Horn, M., Xi, C. and Raskin, L. (2010) Mycobacterium avium infections of Acanthamoeba strains: host strain variability, grazing-acquired infections, and altered dynamics of inactivation with monochloramine. Appl Environ Microbiol 76(19), 6685-6688. Bertram, M.A., Inderlied, C.B., Yadegar, S., Kolanoski, P., Yamada, J. and Young, L. (1986) Confirmation of the beige mouse model for study of disseminated infection with Mycobacterium avium complex. The Journal of Infectious Diseases 154(1), 194-195. Bertrand-Krajewski, J.-L., Chebbo, G. and Saget, A. (1998) Distribution of pollutant mass vs volume in stormwater discharges and the first flush phenomenon. Water Res 32(8), 2341-2356. Beumer, A., King, D., Donohue, M., Mistry, J., Covert, T. and Pfaller, S. (2010) Detection of Mycobacterium avium subsp. paratuberculosis in drinking water and biofilms by quantitative PCR. Applied and Environmental Microbiology 76(21), 7367-7370. Beyrer, K., Lai, S., Dreesman, J., Lee, J., Joseph, C., Harrison, T., Surman-Lee, S., Lück, C., Brodhun, B. and Buchholz, U. (2007) Legionnaires' disease outbreak associated with a cruise liner, August 2003: epidemiological and microbiological findings. Epidemiol Infect 135(05), 802- 810. Bhopal, R. and Fallon, R. (1991) Seasonal variation of Legionnaires' disease in Scotland. Journal of Infection 22(2), 153-160. Birkness, K.A., Swords, W.E., Huang, P.H., White, E.H., Dezzutti, C.S., Lal, R.B. and Quinn, F.D. (1999) Observed differences in virulence-associated phenotypes between a human clinical isolate and a veterinary isolate of Mycobacterium avium. Infection and immunity 67(9), 4895- 4901. Birks, R., Colbourne, J., Hills, S. and Hobson, R. (2004) Microbiological water quality in a large in-building, water recycling facility. Water Science and Technology 50(2), 165-172. Black, W.C. and Berk, S.G. (2003) Cooling towers - A potential environmental source of slow- growing mycobacterial species. Aiha Journal 64(2), 238-242. Blacklock, Z. and Dawson, D. (1979) Atypical mycobacteria causing non-pulmonary disease in Queensland. Pathology 11(2), 283-287. 355 Blanchard, D.C. and Syzdek, L.D. (1982) Water-to-air transfer and enrichment of bacteria in drops from bursting bubbles. Appl Environ Microbiol 43(5), 1001-1005. Bland, C.S., Ireland, J.M., Lozano, E., Alvarez, M.E. and Primm, T.P. (2005) Mycobacterial ecology of the Rio Grande. Applied and Environmental Microbiology 71(10), 5719-5727. Blatny, J.M., Ho, J., Skogan, G., Fykse, E.M., Aarskaug, T. and Waagen, V. (2011) Airborne Legionella bacteria from pulp waste treatment plant: aerosol particles characterized as aggregates and their potential hazard. Aerobiologia 27(2), 147-162. Blatny, J.M., Reif, B.A.P., Skogan, G., Andreassen, O., Høiby, E.A., Ask, E., Waagen, V., Aanonsen, D., Aaberge, I.S. and Caugant, D.A. (2008) Tracking Airborne Legionella and Legionella pneumophila at a Biological Treatment Plant. Environ Sci Technol 42(19), 7360-7367. Boers, T.M. and Ben-Asher, J. (1982) A review of rainwater harvesting. Agricultural Water Management 5(2), 145-158. Bonetta, S., Bonetta, S., Ferretti, E., Balocco, F. and Carraro, E. (2010) Evaluation of Legionella pneumophila contamination in Italian hotel water systems by quantitative real-time PCR and culture methods. J Appl Microbiol 108(5), 1576 - 1583. Boon, N., Top, E.M., Verstraete, W. and Siciliano, S.D. (2003) Bioaugmentation as a tool to protect the structure and function of an activated-sludge microbial community against a 3- chloroaniline shock load. Appl Environ Microbiol 69(3), 1511-1520. Borgen, K., Aaberge, L., Werner-Johansen, O., Gjosund, K., Storsrud, B., Haugsten, S., Nygard, K., Krogh, T., Hoiby, E., Caugant, D., Kanestrom, A., Simonsen, O. and Blystad, H. (2008) Cluster of Legionnaires disease linked to an industrial plant in southeast Norway, June - July 2008. Euro Surveill 13(38). Bouwknegt, M., Schijven, J.F., Schalk, J.A., Husman, R. and Maria, A. (2013) Quantitative risk estimation for a Legionella pneumophila infection due to whirlpool use. Risk Analysis 33(7), 1228-1236. Boyle, D.P., Zembower, T.R., Reddy, S. and Qi, C. (2015) Comparison of clinical features, virulence, and relapse among Mycobacterium avium complex species. American Journal of Respiratory and Critical Care Medicine 191(11), 1310-1317. Breuninger, K.J. and Weir, M.H. (2015) Development of an Interspecies Nested Dose‐Response Model for Mycobacterium avium subspecies paratuberculosis. Risk Analysis 35(8), 1479-1487. 356 Briancesco, R., Semproni, M., Paradiso, R. and Bonadonna, L. (2014) Nontuberculous mycobacteria: an emerging risk in engineered environmental habitats. Annals of Microbiology 64(2), 735-740. Broadhead, A.N., Negron-Alvira, A., Baez, L.A., Hazen, T.C. and Canoy, M.J. (1988) Occurrence of Legionella species in tropical rain water cisterns. Caribbean Journal of Science CRJSA 4 24(1-2). Brodie, E.L., DeSantis, T.Z., Parker, J.P., Zubietta, I.X., Piceno, Y.M. and Andersen, G.L. (2007) Urban aerosols harbor diverse and dynamic bacterial populations. Proc Natl Acad Sci U S A 104(1), 299-304. Brodribb, R., Webster, P. and Farrel, D. (1995) Recurrent Campylobacter fetus subspecies bacteraemia in a febrile neutropaenicpatient linked to tank water. Communicable Disease Intelligence 19(13), 312-313. Brooks, J., Tanner, B., Josephson, K., Gerba, C.P., Haas, C. and Pepper, I.L. (2005a) A national study on the residential impact of biological aerosols from the land application of biosolids. J Appl Microbiol 99(2), 310-322. Brooks, J.P., McLaughlin, M.R., Gerba, C.P. and Pepper, I.L. (2012) Land application of manure and class B biosolids: An occupational and public quantitative microbial risk assessment. Journal of Environmental Quality 41(6), 2009-2023. Brooks, J.P., Tanner, B.D., Gerba, C.P., Haas, C.N. and Pepper, I.L. (2005b) Estimation of bioaerosol risk of infection to residents adjacent to a land applied biosolids site using an empirically derived transport model. J Appl Microbiol 98(2), 397-405. Brooks, R., Parker, B., Gruft, H. and Falkinham 3rd, J. (1984a) Epidemiology of infection by nontuberculous mycobacteria. V. Numbers in eastern United States soils and correlation with soil characteristics. The American review of respiratory disease 130(4), 630-633. Brooks, R.W., George, K.L., Parker, B.C., Falkinham III, J.O. and Gruft, H. (1984b) Recovery and survival of nontuberculous mycobacteria under various growth and decontamination conditions. Canadian journal of microbiology 30(9), 1112-1117. Brooks, R.W., Parker, B.C., Gruft, H. and Falkinham, J.O., 3rd (1984c) Epidemiology of infection by nontuberculous mycobacteria. V. Numbers in eastern United States soils and correlation with soil characteristics. Am Rev Respir Dis 130(4), 630-633. 357 Brown, C.M., Cann, J.W., Simons, G., Fankhauser, R.L., Thomas, W., Parashar, U.D. and Lewis, M.J. (2001) Outbreak of Norwalk virus in a Caribbean island resort: application of molecular diagnostics to ascertain the vehicle of infection. Epidemiol Infect 126(3), 425-432. Brown, J., Sander, S., Craw, D. and Hunter, K. (2005) Measurement of labile metals in acid rock drainage springs, New Zealand: Field application of anodic stripping voltammetry. Applied Geochemistry 20(8), 1533-1545. Brown, S.T., Edwards, F.F., Bernard, E.M., Niki, Y. and Armstrong, D. (1991) Progressive disseminated infection with Mycobacterium avium complex after intravenous and oral challenge in cyclosporine-treated rats. Journal of Infectious Diseases 164(5), 922-927. Brunkard, J.M., Ailes, E., Roberts, V.A., Hill, V., Hilborn, E.D., Craun, G.F., Rajasingham, A., Kahler, A., Garrison, L. and Hicks, L. (2011) Surveillance for waterborne disease outbreaks associated with drinking water—United States, 2007–2008. MMWR Surveill Summ 60(12), 38- 68. Bukh, A.S. and Roslev, P. (2014) Mycobacterium avium complex in day care hot water systems, and persistence of live cells and DNA in hot water pipes. Current microbiology 68(4), 428-439. Buse, H.Y. and Ashbolt, N.J. (2011) Differential growth of Legionella pneumophila strains within a range of amoebae at various temperatures associated with in-premise plumbing. Letters in Applied Microbiology 53(2), 217-224. Buse, H.Y. and Ashbolt, N.J. (2012) Counting Legionella cells within single amoeba host cells. Appl Environ Microbiol 78(6), 2070-2072. Buse, H.Y., Schoen, M.E. and Ashbolt, N.J. (2012) Legionellae in engineered systems and use of quantitative microbial risk assessment to predict exposure. Water Res. Callewaert, C., Van Nevel, S., Kerckhof, F.-M., Granitsiotis, M.S. and Boon, N. (2015) Bacterial Exchange in Household Washing Machines. Frontiers in microbiology 6. Cappelluti, E., Fraire, A.E. and Schaefer, O.P. (2003) A case of" hot tub lung" due to Mycobacterium avium complex in an immunocompetent host. Archives of internal medicine 163(7), 845. Carson, L.A., Bland, L., Cusick, L., Favero, M., Bolan, G., Reingold, A. and Good, R. (1988a) Prevalence of nontuberculous mycobacteria in water supplies of hemodialysis centers. Applied and Environmental Microbiology 54(12), 3122-3125. 358 Carson, L.A., Cusick, L.B., Bland, L. and Favero, M.S. (1988b) Efficacy of chemical dosing methods for isolating nontuberculous mycobacteria from water supplies of dialysis centers. Applied and Environmental Microbiology 54(7), 1756-1760. Carter, R., Holt, M., Pandey, S., Thomson, R., Coulter, C. and Chapman, P. (2014) Epidemiology of Mycobacterium Intracellulare complex in South East Queensland Australia. Am. J. Respir. Crit. Care Med 189, A6537. Casanova, J.L. and Abel, L. (2002) Genetic dissection of immunity to mycobacteria: The human model. Annual Review of Immunology 20(1), 581-620. Cassidy, P.M., Hedberg, K., Saulson, A., McNelly, E. and Winthrop, K.L. (2009) Nontuberculous mycobacterial disease prevalence and risk factors: a changing epidemiology. Clinical Infectious Diseases 49(12), e124-e129. Castilla, J., Barricarte, A., Aldaz, J., Cenoz, M.G., Ferrer, T., Pelaz, C., Pineda, S., Baladrón, B., Martín, I. and Goñi, B. (2008a) A large Legionnaires' disease outbreak in Pamplona, Spain: early detection, rapid control and no case fatality. Epidemiol Infect 136(06), 823-832. Castilla, J., Barricarte, A., Aldaz, J., Garcia, C., Ferrer, T., Pelaz, C., Pineda, S., Baladron, B., Martin, I., Goni, B., Aratajo, P., Chamorro, J., Lameiro, F., Torroba, L., Dorronsoro, L., Martinez-Artola, V., Esparza, M., Gastaminza, M., Fraile, P. and Aldaz, P. (2008b) A large Legionnaires' disease outbreak in Pamplona, Spain: early detection, rapid control and no case fatality. Epidemiol Infect 136(6), 823 - 832. Cayer, M.P., Veillette, M., Pageau, P., Hamelin, R., Bergeron, M.J., Meriaux, A., Cormier, Y. and Duchaine, C. (2007) Identification of mycobacteria in peat moss processing plants: application of molecular biology approaches. Canadian journal of microbiology 53(1), 92-99. CDC (2005) Procedures for the recovery of Legionella from the environment. Division of Bacterial and Mycotic Diseases, R.D.L.S. (ed), Atlanta, GA, USA. CDC (2011) Estimated burden of acute otitis externa--United States, 2003-2007. MMWR. Morbidity and mortality weekly report 60(19), 605. Cerna-Cortes, J.F., Leon-Montes, N., Cortes-Cueto, A.L., Salas-Rangel, L.P., Helguera-Repetto, A.C., Lopez-Hernandez, D., Rivera-Gutierrez, S., Fernandez-Rendon, E. and Gonzalez-y- Merchand, J.A. (2015) Microbiological Quality of Ready-to-Eat Vegetables Collected in Mexico City: Occurrence of Aerobic-Mesophilic Bacteria, Fecal Coliforms, and Potentially Pathogenic Nontuberculous Mycobacteria. BioMed Research International 2015(Article ID 789508). 359 Chan, E.D. and Iseman, M.D. (2010) Slender, Older Women Appear to Be More Susceptible to Nontuberculous Mycobacterial Lung Disease. Gender Medicine 7(1), 5-18. Chan, E.D. and Iseman, M.D. (2013) Underlying host risk factors for nontuberculous mycobacterial lung disease, pp. 110-123. Chang, M., M. McBroom, R. Beasley (2004) Roofing as a source of nonpoint water pollution. Journal of Environmental Management 73(4), 307-315. Chang, Y.C., Le Puil, M., Biggerstaff, J., Randall, A.A., Schulte, A. and Taylor, J.S. (2003) Direct estimation of biofilm density on different pipe material coupons using a specific DNA- probe. Molecular and Cellular Probes 17(5), 237-243. Chapman, H.C., T; Huston, R; O'Toole, J (2008) Water quality and health risks from urban rainwater tanks, Cooperative Research Centre for Water Quality and Treatment, Salisbury SA, Australia. Chen, N.C. and Hanna, S.R. (1978) Drift modeling and monitoring comparisons. Atmospheric Environment (1967) 12(8), 1725-1734. Chern, E.C., King, D., Haugland, R. and Pfaller, S. (2015) Evaluation of quantitative polymerase chain reaction assays targeting Mycobacterium avium, M. intracellulare, and M. avium subspecies paratuberculosis in drinking water biofilms. Chidamba, L. (2015) Microbial Quality of Rainwater Harvested from Rooftops, for Domestic use and Homestead Food Gardens, University of Pretoria South Africa. Chidamba, L. and Korsten, L. (2015) Pyrosequencing analysis of roof-harvested rainwater and river water used for domestic purposes in Luthengele village in the Eastern Cape Province of South Africa. Environ Monit Assess 187(2), 1-17. Chilima, B.Z., Clark, I.M., Floyd, S., Fine, P.E. and Hirsch, P.R. (2006) Distribution of environmental mycobacteria in Karonga District, northern Malawi. Applied and Environmental Microbiology 72(4), 2343-2350. Chinsembu, K.C. and Hakwenye, H. (2010) Isolation and serotyping of Legionella pneumophila from Goreangab Dam and hostel shower heads in Namibia. African Journal of Microbiology Research 4(18), 1890-1896. 360 Cirillo, J.D., Falkow, S., Tompkins, L.S. and Bermudez, L.E. (1997) Interaction of Mycobacterium avium with environmental amoebae enhances virulence. Infection and immunity 65(9), 3759-3767. Clark, S.E., Steele, K.A., Spicher, J., Siu, C.Y., Lalor, M.M., Pitt, R. and Kirby, J.T. (2008) Roofing materials’ contributions to storm-water runoff pollution. Journal of Irrigation and Drainage Engineering 134(5), 638-645. Cochran, W.G. (1954) The combination of estimates from different experiments. Biometrics 10(1), 101-129. Coetzee, N., Duggal, H., Hawker, J., Ibbotson, S., Harrison, T., Phin, N., Laza-Stanca, V., Johnston, R., Iqbal, Z. and Rehman, Y. (2012) An outbreak of Legionnaires’ disease associated with a display spa pool in retail premises, Stoke-on-Trent, United Kingdom, July 2012. Euro Surveill 17(37), 1-4. Collins, F.M. (1985) Protection to mice afforded by BCG vaccines against an aerogenic challenge by three mycobacteria of decreasing virulence. Tubercle 66(4), 267-276. Collins, F.M., Morrison, N. and Montalbine, V. (1978) Immune response to persistent mycobacterial infection in mice. Infection and immunity 20(2), 430-438. Collins, F.M. and Watson, S.R. (1981) Immune responses to atypical mycobacterial lung infections. Review of Infectious Diseases 3(5), 981-989. Collins, P., Matthews, P., McDiarmid, A. and BROWN, A. (1983) The pathogenicity of Mycobacterium avium and related mycobacteria for experimental animals. Journal of medical microbiology 16(1), 27-35. Cook, K., Flis, S. and Ballard, C. (2013) Sensitivity of Mycobacterium avium subsp paratuberculosis, Escherichia coli and Salmonella enterica serotype Typhimurium to low pH, high organic acids and ensiling. Journal of applied microbiology 115(2), 334-345. Coombes, P.J. (2002) Rainwater tanks revisited: new opportunities for urban water cycle management. PhD thesis, University of Newcastle, Newcastle, Australia. Coombes, P.J., Argue, J.R. and Kuczera, G. (1999) Figtree Place: A case study in water sensitive urban development (WSUD). Urban Water 1(4), 335-343. 361 Cooney, R., Kazda, J., Quinn, J., Cook, B., Muller, K. and Monaghan, M. (1997) Environmental mycobacteria in Ireland as a source of non-specific sensitisation to tuberculins. Irish Veterinary Journal 50(6), 370-373. Cooper, I. and Hanlon, G. (2010) Resistance of Legionella pneumophila serotype 1 biofilms to chlorine-based disinfection. Journal of Hospital Infection 74(2), 152-159. Costallat, L., De Castro, P., Rosdrigues, A. and Rosdrigues, F. (1977) Examination of soils in the Campinas rural area for microorganisms of the Mycobacterium avium-intracellulare- scrofulaceum complex. Australian veterinary journal 53(7), 349-350. Covert, T.C., Rodgers, M.R., Reyes, A.L. and Stelma, G.N. (1999) Occurrence of nontuberculous mycobacteria in environmental samples. Applied and Environmental Microbiology 65(6), 2492- 2496. Crabtree, K.R., RH; Shaw, SB; Rose, JB (1996) The detection of cryptosporidium oocysts and giardia cysts in cistern water in the US Virgin Islands. Water Res 30(1), 208-216. Crago, B., Ferrato, C., Drews, S.J., Louie, T., Ceri, H., Turner, R.J., Roles, A. and Louie, M. (2014) Surveillance and molecular characterization of non-tuberculous mycobacteria in a hospital water distribution system over a three-year period. J Hosp Infect 87(1), 59-62. Craun, G.F., Brunkard, J.M., Yoder, J.S., Roberts, V.A., Carpenter, J., Wade, T., Calderon, R.L., Roberts, J.M., Beach, M.J. and Roy, S.L. (2010) Causes of outbreaks associated with drinking water in the United States from 1971 to 2006. Clinical Microbiology Reviews 23(3), 507-528. Cross, M., Labes, R. and Mackintosh, C. (2000) Oral Infection of Ferrets with Virulent< i> Mycobacterium bovis or< i> Mycobacterium avium: Susceptibility, Pathogenesis and Immune Response. Journal of Comparative Pathology 123(1), 15-21. Crowle, A., Ross, E., Cohn, D., Gilden, J. and May, M. (1992) Comparison of the abilities of Mycobacterium avium and Mycobacterium intracellulare to infect and multiply in cultured human macrophages from normal and human immunodeficiency virus-infected subjects. Infection and immunity 60(9), 3697-3703. Crowley, B.J. (2005) Neighborhood level analysis of rainwater catchment in Portland, OR. Master's, Portland State University. Cvetnic, Z., Kovacic, H. and Ocepek, M. (1998) Mycobacteria in the environment and in the feed of swine in Croatia. Wiener Tierarztliche Monatsschrift 85(1), 18-21. 362 Daley, C. and Griffith, D. (2010) Pulmonary non-tuberculous mycobacterial infections. International Journal of Tuberculosis and Lung Disease 14(6), 665-671. Dawson, D. (1971) Potential pathogens among strains of mycobacteria isolated from house-dusts. Med J Aust 1. De Groote, M.A., Pace, N.R., Fulton, K. and Falkinham, J.O., 3rd (2006) Relationships between Mycobacterium isolates from patients with pulmonary mycobacterial infection and potting soils. Appl Environ Microbiol 72(12), 7602-7606. de Kwaadsteniet, M., Dobrowsky, P., van Deventer, A., Khan, W. and Cloete, T. (2013) Domestic rainwater harvesting: microbial and chemical water quality and point-of-use treatment systems. Water, Air, & Soil Pollution 224(7), 1-19. de Man, H., Bouwknegt, M., van Heijnsbergen, E., Leenen, E.J., van Knapen, F. and de Roda Husman, A.M. (2014a) Health risk assessment for splash parks that use rainwater as source water. Water Res 54, 254-261. de Man, H., Heederik, D.D.J., Leenen, E.J.T.M., de Roda Husman, A.M., Spithoven, J.J.G. and van Knapen, F. (2014b) Human exposure to endotoxins and fecal indicators originating from water features. Water Res (0). de Man, H., van den Berg, H., Leenen, E., Schijven, J., Schets, F., van der Vliet, J., van Knapen, F. and de Roda Husman, A. (2014c) Quantitative assessment of infection risk from exposure to waterborne pathogens in urban floodwater. Water Res 48, 90-99. Dean, J. and Hunter, P.R. (2012) Risk of gastrointestinal illness associated with the consumption of rainwater: a systematic review. Environ Sci Technol 46(5), 2501-2507. Delafont, V., Mougari, F., Cambau, E., Joyeux, M., Bouchon, D., Héchard, Y. and Moulin, L. (2014) First evidence of amoebae–mycobacteria association in drinking water network. Environ Sci Technol 48(20), 11872-11882. Dennis, P. and Lee, J. (1988) Differences in aerosol survival between pathogenic and non‐ pathogenic strains of Legionella pneumophila serogroup 1. Journal of Applied Bacteriology 65(2), 135-141. Dicaprio, E., Ma, Y., Purgianto, A., Hughes, J. and Li, J. (2012) Internalization and dissemination of human norovirus and animal caliciviruses in hydroponically grown romaine lettuce. Appl Environ Microbiol 78(17), 6143-6152. 363 Diederen, B. (2008) Legionella spp. and Legionnaires' disease. Journal of Infection 56(1), 1-12. Dirac, M.A., Horan, K.L., Doody, D.R., Meschke, J.S., Park, D.R. and Jackson, L.A. (2012) Environment or host?: A case–control study of risk factors for Mycobacterium avium complex lung disease. Am J Respir Crit Care Med 186. Dobrowsky, P., Carstens, M., De Villiers, J., Cloete, T. and Khan, W. (2015a) Efficiency of a closed-coupled solar pasteurization system in treating roof harvested rainwater. Science of The Total Environment 536, 206-214. Dobrowsky, P., Lombard, M., Cloete, W., Saayman, M., Cloete, T., Carstens, M., Khan, S. and Khan, W. (2015b) Efficiency of microfiltration systems for the removal of bacterial and viral contaminants from surface and rainwater. Water, Air, & Soil Pollution 226(3), 1-14. Dobrowsky, P.H., De Kwaadsteniet, M., Cloete, T.E. and Khan, W. (2014) Distribution of indigenous bacterial pathogens and potential pathogens associated with roof-harvested rainwater. Appl Environ Microbiol 80(7), 2307-2316. Domènech, L. and Saurí, D. (2011) A comparative appraisal of the use of rainwater harvesting in single and multi-family buildings of the Metropolitan Area of Barcelona (Spain): social experience, drinking water savings and economic costs. Journal of Cleaner Production 19(6), 598-608. Donohue, M.J., Mistry, J.H., Donohue, J.M., O’Connell, K., King, D., Byran, J., Covert, T. and Pfaller, S. (2015) Increased frequency of nontuberculous mycobacteria detection at potable water taps within the United States. Environmental science & technology 49(10), 6127-6133. Doust, R., Mobares, A., Mirkalantar, S., Aslani, J. and Fuladi, A. (2009) Nosocomial Legionnaires' disease outbreak in Tehran. Research Journal of Microbiology 4(1), 23-30. Dowd, S.E., Gerba, C.P., Pepper, I.L. and Pillai, S.D. (2000) Bioaerosol transport modeling and risk assessment in relation to biosolid placement. Journal of Environmental Quality 29(1), 343- 348. Doyle, K.C. (2008) Sizing the first flush and its effect on the storage-reliability-yield behavior of rainwater harvesting in Rwanda, Citeseer. Drancourt, M. (2014) Looking in amoebae as a source of mycobacteria. Microbial pathogenesis (In Press). 364 Du Moulin, G. and Stottmeier, K. (1986) Waterborne mycobacteria: An increasing threat to health. ASM American Society for Microbiology News 52(10), 525-529. du Moulin, G.C., Sherman, I.H., Hoaglin, D.C. and Stottmeier, K.D. (1985) Mycobacterium avium complex, an emerging pathogen in Massachusetts. Journal of Clinical Microbiology 22. du Moulin, G.C., Stottmeier, K.D., Pelletier, P.A., Tsang, A.Y. and Hedley-Whyte, J. (1988) Concentration of Mycobacterium avium by hospital hot water systems. Jama 260(11), 1599-1601. Dufour, A.P., Evans, O., Behymer, T.D. and Cantu, R. (2006) Water ingestion during swimming activities in a pool: a pilot study. Journal of Water and Health 4(4), 425-430. Dungan, R. (2010) Board-invited review: Fate and transport of bioaerosols associated with livestock operations and manures. Journal of animal science 88(11), 3693-3706. Dungan, R.S. (2014) Estimation of infectious risks in residential populations exposed to airborne pathogens during center pivot irrigation of dairy wastewaters. Environ Sci Technol 48(9), 5033- 5042. Dvorska, L., Bartos, M., Ostadal, O., Kaustova, J., Matlova, L. and Pavlik, I. (2002) IS1311 and IS1245 restriction fragment length polymorphism analyses, serotypes, and drug susceptibilities of Mycobacterium avium complex isolates obtained from a human immunodeficiency virus-negative patient. Journal of clinical microbiology 40(10), 3712-3719. Eaton, T., Falkinham 3rd, J. and von Reyn, C.F. (1995a) Recovery of Mycobacterium avium from cigarettes. Journal of clinical microbiology 33(10), 2757. Eaton, T., Falkinham, J.O., Aisu, T.O. and Daniel, T.M. (1995b) Isolation and characteristics of Mycobacterium avium complex from water and soil samples in Uganda. Tuber Lung Dis 76(6), 570-574. Eberhart-Phillips, J., Walker, N., Garrett, N., Bell, D., Sinclair, D., Rainger, W. and Bates, M. (1997) Campylobacteriosis in New Zealand: results of a case-control study. Journal of Epidemiology and Community Health 51(6), 686-691. Edwards, L.B., Acquaviva, F.A., Livesay, V.T., Cross, F.W. and Palmer, C.E. (1969) An atlas of sensitivity to tuberculin, PPD-B, and histoplasmin in the United States. Am Rev Respir Dis 99(4), Suppl:1-132. 365 Egodawatta, P., Thomas, E. and Goonetilleke, A. (2009) Understanding the physical processes of pollutant build-up and wash-off on roof surfaces. Science of The Total Environment 407(6), 1834-1841. Ellis, K. (1993) Legionellosis: A concise review. Water and Environment Journal 7(4), 418-430. Ellsworth, S., Kirkbride, C. and Johnson, D. (1980) Excretion of Mycobacaterium avium from lesions in the intestine and tonsils of infected swine. American journal of veterinary research 41(9), 1526. Embil, J., Warren, P., Yakrus, M., Stark, R., Corne, S., Forrest, D. and Hershfield, E. (1997) Pulmonary illness associated with exposure to Mycobacterium-avium complex in hot tub water. Chest 111(3), 813-816. Engelhart, S., Pleischl, S., Lück, C., Marklein, G., Fischnaller, E., Martin, S., Simon, A. and Exner, M. (2008) Hospital-acquired legionellosis originating from a cooling tower during a period of thermal inversion. Int J Hyg Environ Health 211(3), 235-240. enHealth (2004) Guidance on the use of rainwater tanks, Australian Government Department of Health and Ageing, Canberra, Australia. Euser, S., Bruin, J., Mooi-Kokenberg, E., Peeters, M., Verbakel, H., Yzerman, E. and Den Boer, J. (2012) Diagnostic testing for Legionnaires’ disease in the Netherlands between 2007 and 2009: a possible cause for the decline in reported Legionnaires’ disease patients. European journal of clinical microbiology & infectious diseases 31(8), 1969-1974. Euser, S.M., Pelgrim, M. and Den Boer, J.W. (2010) Legionnaires' disease and Pontiac fever after using a private outdoor whirlpool spa. Scandinavian journal of infectious diseases 42(11-12), 910-916. Evans, C.A., Coombes, P.J. and Dunstan, R.H. (2006) Wind, rain and bacteria: The effect of weather on the microbial composition of roof-harvested rainwater. Water Res 40(1), 37-44. Evans, C.A., Coombes, P.J., Dunstan, R.H. and Harrison, T. (2009) Extensive bacterial diversity indicates the potential operation of a dynamic micro-ecology within domestic rainwater storage systems. Sci Total Environ 407(19), 5206-5215. Everman, J.L., Ziaie, N.R., Bechler, J. and Bermudez, L.E. (2015) Establishing Caenorhabditis elegans as a model for Mycobacterium avium subspecies hominissuis infection and intestinal colonization. Biology open 4(10), 1330-1335. 366 Fairchok, M.P., Rouse, J.H. and Morris, S.L. (1995) Age-dependent humoral responses of children to mycobacterial antigens. Clinical and diagnostic laboratory immunology 2(4), 443-447. Falkinham 3rd, J., Hilborn, E.D., Arduino, M.J., Pruden, A. and Edwards, M.A. (2015) Epidemiology and Ecology of Opportunistic Premise Plumbing Pathogens: Legionella pneumophila, Mycobacterium avium, and Pseudomonas aeruginosa. Environ Health Perspect 123(8), 749-758. Falkinham III, J., Iseman, M., Haas, P. and Soolingen, D. (2008) Mycobacterium avium in a shower linked to pulmonary disease. Journal of water and health 6(2), 209-213. Falkinham III, J., Parker, B. and Gruft, H. (1980) Epidemiology of infection by nontuberculous mycobacteria. I. Geographic distribution in the eastern United States. The American review of respiratory disease 121(6), 931-937. Falkinham III, J.O. (1996) Epidemiology of infection by nontuberculous mycobacteria. Clinical Microbiology Reviews 9(2), 177. Falkinham III, J.O. (1998) Mycobacteria, pp. 178-209, Springer. Falkinham III, J.O. (2003) Factors influencing the chlorine susceptibility of Mycobacterium avium, Mycobacterium intracellulare, and Mycobacterium scrofulaceum. Applied and Environmental Microbiology 69(9), 5685-5689. Falkinham III, J.O. (2011) Nontuberculous mycobacteria from household plumbing of patients with nontuberculous mycobacteria disease. Emerging infectious diseases 17(3), 419. Falkinham III, J.O. (2013a) Ecology of Nontuberculous Mycobacteria-Where Do Human Infections Come from? Seminars in Respiratory and Critical Care Medicine 34(1), 95-102. Falkinham III, J.O. (2013b) Ecology of Nontuberculous Mycobacteria—Where Do Human Infections Come from?, pp. 095-102, Thieme Medical Publishers. Falkinham III, J.O. (2016) Current Epidemiologic Trends of the Nontuberculous Mycobacteria (NTM). Current environmental health reports 3(2), 161-167. Falkinham III, J.O., Norton, C.D. and LeChevallier, M.W. (2001) Factors influencing numbers of Mycobacterium avium, Mycobacterium intracellulare, and other Mycobacteria in Drinking Water Distribution Systems. Applied and Environmental Microbiology 67(3), 1225-1231. 367 Falkinham, J., Norton, C. and Le Chavallier, M. (2001a) Factors influencing numbers of Mycobacterium avium, Mycobacterium intracellulare and other mycobacteria in drinking water distribution systems. Appl Environ Microbiol 67. Falkinham, J.O. (2003) Mycobacterial aerosols and respiratory disease. Emerging infectious diseases 9(7), 763-767. Falkinham, J.O. (2010) Hospital water filters as a source of Mycobacterium avium complex. Journal of medical microbiology 59(10), 1198-1202. Falkinham, J.O. (2013) Ecology of Nontuberculous Mycobacteria—Where Do Human Infections Come from? Semin Respir Crit Care Med 34(01), 095-102. Falkinham, J.O., 3rd (2011) Nontuberculous mycobacteria from household plumbing of patients with nontuberculous mycobacteria disease. Emerg Infect Dis 17(3), 419-424. Falkinham, J.O., Norton, C.D. and LeChevallier, M.W. (2001b) Factors Influencing Numbers of Mycobacterium avium, Mycobacterium intracellulare, and Other Mycobacteria in Drinking Water Distribution Systems. Appl Environ Microbiol 67(3), 1225-1231. Falkinham, J.O., Pruden, A. and Edwards, M. (2015) Opportunistic Premise Plumbing Pathogens: Increasingly Important Pathogens in Drinking Water. Pathogens 4(2), 373-386. Farhat, M., Greenaway, C., Pai, M. and Menzies, D. (2006) False-positive tuberculin skin tests: what is the absolute effect of BCG and non-tuberculous mycobacteria? The International Journal of Tuberculosis and Lung Disease 10(11), 1192-1204. Farreny, R., Morales-Pinzon, T., Guisasola, A., Taya, C., Rieradevall, J. and Gabarrell, X. (2011) Roof selection for rainwater harvesting: quantity and quality assessments in Spain. Water Res 45(10), 3245-3254. Fay, M.P. and Shaw, P.A. (2010) Exact and Asymptotic Weighted Logrank Tests for Interval Censored Data: The interval R Package. 2010 36(2), 34. Feazel, L., Baumgartner, L., Peterson, K., Frank, D., Harris, J. and Pace, N. (2009) Opportunistic pathogens enriched in showerhead biofilms. PNAS 106. Ferre, M., Arias, C., Oliva, J., Pedrol, A., Garcia, M., Pellicer, T., Roura, P. and Dominguez, A. (2009) A community outbreak of Legionnaires' disease associated with a cooling tower in Vic and Gurb, Catalonia (Spain) in 2005. Eur J Clin Microbiol Infect Dis 28(2), 153 - 159. 368 Fewtrell, L. and Kay, D. (2007a) Microbial quality of rainwater supplies in developed countries: A review. Urban Water Journal 4(4), 253-260. Fewtrell, L. and Kay, D. (2007b) Quantitative microbial risk assessment with respect to Campylobacter spp. in toilets flushed with harvested rainwater. Water and Environment Journal 21(4), 275-280. Field, S.K., Fisher, D. and Cowie, R.L. (2004) Mycobacterium avium complex pulmonary disease in patients without HIV infection. CHEST Journal 126(2), 566-581. Fipps, G. (2003) Irrigation Water Quality Standards and Salinity Management Strategies, pp. pp. 1-20, AgriLife Extension, Texas Cooperative Extension B-1667, 4-03, Texas A&M University, College Station. Fisman, D.L., S; Wellenius, GA; Johnson, C; Britz, P; Gaskins, M; Maher, J; Mittleman, MA; Spain, V; Haas, CN; Newbern, C (2005) It's not the heat, it's the humidity: Wet weather increases legionellosis risk in the greater Philadelphia metropolitan area. Journal of Infectious Diseases 192, 2066-2073. Fisman, D.N. (2007) Seasonality of infectious diseases. Annu. Rev. Public Health 28, 127-143. Fjällbrant, H., Akerstrom, M., Svensson, E. and Andersson, E. (2013) Hot tub lung: an occupational hazard. European Respiratory Review 22(127), 88-90. Fleischman, R., Du Moulin, G., Esber, H., Ilievski, V. and Bogden, A. (1982) Nontuberculous mycobacterial infection attributable to Mycobacterium intracellulare serotype 10 in two rhesus monkeys. Journal of the American Veterinary Medical Association 181(11), 1358-1362. Fliermans, C. (1996) Ecology of Legionella: from data to knowledge with a little wisdom. Microbial Ecology 32(2), 203-228. Fogarty, L.R., Haack, S.K., Johnson, H.E., Brennan, A.K., Isaacs, N.M. and Spencer, C. (2015) Staphylococcus aureus and methicillin-resistant S. aureus (MRSA) at ambient freshwater beaches. Journal of Water and Health 13(3), 680-692. Forster, J. (1996) Patterns of roof runoff contamination and their potential implications on practice and regulation of treatment and local infiltration, pp. 39-48, Pergamon-Elsevier Science Ltd. Forster, J. (1999) Variability of roof runoff quality. Water Science and Technology 39(5), 137- 144. 369 Fossum, H., Reif, B.P., Tutkun, M. and Gjesdal, T. (2012) On the use of computational fluid dynamics to investigate aerosol dispersion in an industrial environment: a case study. Boundary- layer meteorology 144(1), 21-40. Franklin, L.J., Fielding, J.E., Gregory, J., Gullan, L., Lightfoot, D., Poznanski, S.Y. and Vally, H. (2009) An outbreak of Salmonella Typhimurium 9 at a school camp linked to contamination of rainwater tanks. Epidemiol Infect 137(3), 434-440. Fujita, K., Ito, Y., Hirai, T., Kubo, T., Maekawa, K., Togashi, K., Ichiyama, S. and Mishima, M. (2014) Association between polyclonal and mixed mycobacterial Mycobacterium avium complex infection and environmental exposure. Ann Am Thorac Soc 11(1), 45-53. Fujita, K., Ito, Y., Hirai, T., Maekawa, K., Imai, S., Tatsumi, S., Niimi, A., Iinuma, Y., Ichiyama, S. and Mishima, M. (2013) Genetic relatedness of Mycobacterium avium‐intracellulare complex isolates from patients with pulmonary MAC disease and their residential soils. Clinical Microbiology and Infection 19(6), 537-541. Galada, H., Gurian, P., Joe, A., Kumar, A., Olson, B., Olson, M., Richter, E., Teng, J., Zhang, H. and Xagoraraki, I. (2012) Site Specific Risk Assessment Tool for Land Applied Biosolids. Water Environment Research Foundation: Alexandria, VA, USA. Gale, P. (2005) Land application of treated sewage sludge: quantifying pathogen risks from consumption of crops. J Appl Microbiol 98(2), 380-396. Gangadharam, P., Edwards 3rd, C., Murthy, P. and Pratt, P. (1983) An acute infection model for Mycobacterium intracellulare disease using beige mice: preliminary results. The American review of respiratory disease 127(5), 648. García-Fulgueiras, A., Navarro, C., Fenoll, D., García, J., González-Diego, P., Jiménez-Buñuales, T., Rodriguez, M., Lopez, R., Pacheco, F. and Ruiz, J. (2003) Legionnaires' disease outbreak in Murcia, Spain. Emerging Infectious Diseases 9(8), 915-921. Gaynor, W.T., Cousins, D.V. and Friend, J.A. (1990) Mycobacterial infection in numbats (Myrmecobius fasciatus). Journal of Zoo and Wildlife Medicine, 476-479. George, K.L., Parker, B.C., Gruft, H. and Falkinham III, J.O. (1980) Epidemiology of Infection by Nontuberculous Mycobacteria: II. Growth and Survival in Natural Waters 1, 2. American Review of Respiratory Disease 122(1), 89-94. Gerba, C.P. and Kennedy, D. (2007) Enteric virus survival during household laundering and impact of disinfection with sodium hypochlorite. Appl Environ Microbiol 73(14), 4425-4428. 370 Gibson, L.L., Rose, J.B. and Haas, C.N. (1999) Use of quantitative microbial risk assessment for evaluation of the benefits of laundry sanitation. American journal of infection control 27(6), S34- S39. Gikas, G.D. and Tsihrintzis, V.A. (2012) Assessment of water quality of first-flush roof runoff and harvested rainwater. Journal of Hydrology 466, 115-126. Gilmour, M.W., Bernard, K., Tracz, D.M., Olson, A.B., Corbett, C.R., Burdz, T., Ng, B., Wiebe, D., Broukhanski, G. and Boleszczuk, P. (2007) Molecular typing of a Legionella pneumophila outbreak in Ontario, Canada. Journal of medical microbiology 56(3), 336-341. Glazer, C., Martyny, J. and Rose, C. (2008) Hot tub associated granulomatous lung disease from mycobacterial bioaerosols. Clinical Pulmonary Medicine 15(3), 138-144. Glazer, C.S., Martyny, J.W., Lee, B., Sanchez, T.L., Sells, T.M., Newman, L.S., Murphy, J., Heifets, L. and Rose, C.S. (2007) Nontuberculous mycobacteria in aerosol droplets and bulk water samples from therapy pools and hot tubs. Journal of Occupational and Environmental Hygiene 4(11), 831-840. Glover, N., Holtzman, A., Aronson, T., Froman, S., Berlin, O., Dominguez, P., Kunkel, K., Overturf, G., Stelma Jr, G. and Smith, C. (1994) The isolation and identification of Mycobacterium avium complex (MAC) recovered from Los Angeles potable water, a possible source of infection in AIDS patients. International Journal of Environmental Health Research 4(2), 63-72. Gobel, P., Dierkes, C. and Coldewey, W.G. (2007) Storm water runoff concentration matrix for urban areas. J Contam Hydrol 91(1-2), 26-42. Goldstein, R.E.R., Micallef, S.A., Gibbs, S.G., He, X., George, A., Sapkota, A., Joseph, S.W. and Sapkota, A.R. (2014) Occupational exposure to Staphylococcus aureus and Enterococcus spp. among spray irrigation workers using reclaimed water. International journal of environmental research and public health 11(4), 4340-4355. Goslee, S. and Wolinsky, E. (1976) Water as a Source of Potentially Pathogenic Mycobacteria 1– 4. American Review of Respiratory Disease 113(3), 287-292. Grabow, N., Pienaar, E. and Kfir, R. (1991) The occurrence of Legionella bacteria in cooling towers in South Africa. Water Science and Technology 24(2), 149-152. Graham, L., Warren, N., Tsang, A. and Dalton, H. (1988) Mycobacterium avium complex pseudobacteriuria from a hospital water supply. Journal of clinical microbiology 26(5), 1034- 1036. 371 Griffith, D.E., Aksamit, T., Brown-Elliott, B.A., Catanzaro, A., Daley, C. and Gordin, F. (2007) An official ATS/IDSA statement: diagnosis, treatment, and prevention of nontuberculous mycobacterial diseases. Am J Respir Crit Care Med 175. Griffith, D.E., Brown-Elliott, B.A. and Wallace Jr, R.J. (2015) Hit the Road, MAC, and Don’t You Come Back No More. American Journal of Respiratory and Critical Care Medicine 191(11), 1222-1224. Gross, W., Hawkins, J. and Murphy, D. (1975) Origin and significance of Mycobacterium xenopi in clinical specimens. I. Water as a source of contamination. Bulletin of the International Union against Tuberculosis 51(1), 267-269. Gruft, H., Falkinham, J.O. and Parker, B.C. (1981) Recent experience in the epidemiology of disease caused by atypical mycobacteria. Review of Infectious Diseases 3(5), 990-996. Gubler, J., Salfinger, M. and von Graevenitz, A. (1992) Pseudoepidemic of nontuberculous mycobacteria due to a contaminated bronchoscope cleaning machine. Report of an outbreak and review of the literature. CHEST Journal 101(5), 1245-1249. Gudiol, C., Verdaguer, R., Angeles Domínguez, M., Fernández‐Sevilla, A. and Carratalà, J. (2007) Outbreak of Legionnaires' disease in immunosuppressed patients at a cancer centre: usefulness of universal urine antigen testing and early levofloxacin therapy. Clinical Microbiology and Infection 13(11), 1125-1128. Guthertz, L.S., Damsker, B., Bottone, E.J., Ford, E.G., Midura, T.F. and Janda, J.M. (1989) Mycobacterium avium and Mycobacterium intracellulare infections in patients with and without AIDS. Journal of Infectious Diseases 160(6), 1037-1041. Haas, C.N. (2015) Microbial Dose Response Modeling: Past, Present, and Future. Environ Sci Technol 49(3), 1245-1259. Haas, C.N., Rose, J.B. and Gerba, C.P. (1999) Quantitative microbial risk assessment, Wiley. Haas, C.N., Rose, J.B. and Gerba, C.P. (2014) Quantitative Microbial Risk Assessment, pp. i-xii, John Wiley & Sons, Inc. Halstrom, S., Price, P. and Thomson, R. (2015) Review: Environmental mycobacteria as a cause of human infection. International Journal of Mycobacteriology 4(2), 81-91. Hambleton, P., Broster, M., Dennis, P., Henstridge, R., Fitzgeorge, R. and Conlan, J. (1983) Survival of virulent Legionella pneumophila in aerosols. Journal of Hygiene 90(03), 451-460. 372 Hamilton, K., Ahmed, W., Palmer, A., Sidhu, J., Hodgers, L., Toze, S. and Haas, C. (2016) Public health implications of Acanthamoeba and multiple potential opportunistic pathogens in roof-harvested rainwater tanks. Environmental Research 150, 320-327. Hamilton, K.A. and Haas, C.N. (2016) Critical review of mathematical approaches for quantitative microbial risk assessment (QMRA) of Legionella in engineered water systems: research gaps and a new framework. Environmental Science: Water Research & Technology 2(4), 599-613. Hamilton, K.A., Weir, M.H. and Haas, C.N. (2017) Dose response models and a quantitative microbial risk assessment framework for the Mycobacterium avium complex that account for recent developments in molecular biology, taxonomy, and epidemiology. Water Research 109, 310-326. Han, X.Y., Tarrand, J.J., Infante, R., Jacobson, K.L. and Truong, M. (2005) Clinical significance and epidemiologic analyses of Mycobacterium avium and Mycobacterium intracellulare among patients without AIDS. Journal of clinical microbiology 43(9), 4407-4412. Hanak, V., Kalra, S., Aksamit, T.R., Hartman, T.E., Tazelaar, H.D. and Ryu, J.H. (2006) Hot tub lung: Presenting features and clinical course of 21 patients. Respiratory Medicine 100(4), 610- 615. Handia, L. (2005) Comparative study of rainwater quality in urban Zambia. Journal of Water Supply Research and Technology-Aqua 54(1), 55-64. Hanjra, M.A., Blackwell, J., Carr, G., Zhang, F. and Jackson, T.M. (2012) Wastewater irrigation and environmental health: Implications for water governance and public policy. Int J Hyg Environ Health 215(3), 255-269. Harwood, V.J., Levine, A.D., Scott, T.M., Chivukula, V., Lukasik, J., Farrah, S.R. and Rose, J.B. (2005) Validity of the indicator organism paradigm for pathogen reduction in reclaimed water and public health protection. Appl Environ Microbiol 71(6), 3163-3170. Hathaway, S.C. and Cook, R.L. (1997) A regulatory perspective on the potential uses of microbial risk assessment in international trade. International Journal of Food Microbiology 36(2), 127-133. Hatt, B., Deletic, A. and Fletcher, T. (2004) Integrated stormwater treatment and re-use systems: Inventory of Australian practice, CRC for Catchment Hydrology. Hauer, T. (2010) Phototrophic biofilms on the interior walls of concrete Iterson-type cooling towers. Journal of applied phycology 22(6), 733-736. 373 Haugland, R.A., Siefring, S.C., Wymer, L.J., Brenner, K.P. and Dufour, A.P. (2005) Comparison of Enterococcus measurements in freshwater at two recreational beaches by quantitative polymerase chain reaction and membrane filter culture analysis. Water Res 39(4), 559-568. Haupt, T.E., Heffernan, R.T., Kazmierczak, J.J., Nehls-Lowe, H., Rheineck, B., Powell, C., Leonhardt, K.K., Chitnis, A.S. and Davis, J.P. (2012) An outbreak of Legionnaires disease associated with a decorative water wall fountain in a hospital. Infection Control 33(02), 185-191. Havelaar, A., Berwald, L., Groothuis, D. and Baas, J. (1985) Mycobacteria in semi-public swimming-pools and whirlpools. Zentralblatt für Bakteriologie, Mikrobiologie und Hygiene. 1. Abt. Originale B, Hygiene 180(5-6), 505-514. Hazra, R., Lee, S.H., Maslow, J.N. and Husson, R.N. (2000) Related strains of Mycobacterium avium cause disease in children with AIDS and in children with lymphadenitis. Journal of Infectious Diseases 181(4), 1298-1303. Helsel, D.R. (2011) Statistics for censored environmental data using Minitab and R, John Wiley & Sons. Helsel, D.R. and Hirsch, R.M. (2002) Statistical methods in water resources, US Geological survey Reston, VA. Hendricks, E.E., Ludlage, E., Bussell, S., George, K., Wegner, F.H. and Mansfield, K.G. (2004) Wasting syndrome and disruption of the somatotropic axis in simian immunodeficiency virus- infected macaques with Mycobacterium avium complex infection. Journal of Infectious Diseases 190(12), 2187-2194. Heyder, J., Gebhart, J., Rudolf, G., Schiller, C.F. and Stahlhofen, W. (1986) Deposition of particles in the human respiratory tract in the size range 0.005–15 μm. Journal of Aerosol Science 17(5), 811-825. Heyworth, J.S., Glonek, G., Maynard, E.J., Baghurst, P.A. and Finlay-Jones, J. (2006) Consumption of untreated tank rainwater and gastroenteritis among young children in South Australia. Int J Epidemiol 35(4), 1051-1058. Hibiya, K., Furugen, M., Higa, F., Tateyama, M. and Fujita, J. (2011) Pigs as an experimental model for systemic Mycobacterium avium infectious disease. Comparative immunology, microbiology and infectious diseases 34(6), 455-464. Hicks, L., Rose, C., Fields, B., Drees, M., Engel, J., Jenkins, P., Rouse, B., Blythe, D., Khalifah, A. and Feikin, D. (2007) Increased rainfall is associated with increased risk for legionellosis. Epidemiol Infect 135(5), 811-817. 374 Hilborn, E.D., Covert, T.C., Yakrus, M.A., Harris, S.I., Donnelly, S.F., Rice, E.W., Toney, S., Bailey, S.A. and Stelma, G.N. (2006) Persistence of nontuberculous mycobacteria in a drinking water system after addition of filtration treatment. Applied and Environmental Microbiology 72(9), 5864-5869. Hilborn, E.D., Yakrus, M.A., Covert, T.C., Harris, S.I., Donnelly, S.F., Schmitt, M.T., Toney, S., Bailey, S.A. and Stelma, G.N. (2008) Molecular comparison of Mycobacterium avium isolates from clinical and environmental sources. Applied and Environmental Microbiology 74(15), 4966- 4968. Hill, V.R., Kahler, A.M., Jothikumar, N., Johnson, T.B., Hahn, D. and Cromeans, T.L. (2007) Multistate evaluation of an ultrafiltration-based procedure for simultaneous recovery of enteric microbes in 100-liter tap water samples. Appl Environ Microbiol 73(13), 4218-4225. Hill, V.R., Polaczyk, A.L., Hahn, D., Narayanan, J., Cromeans, T.L., Roberts, J.M. and Amburgey, J.E. (2005) Development of a rapid method for simultaneous recovery of diverse microbes in drinking water by ultrafiltration with sodium polyphosphate and surfactants. Appl Environ Microbiol 71(11), 6878-6884. Hinds, W.C. (1999) Aerosol technology: properties, behavior, and measurement of airborne particles, Wiley-Interscience, New York. Hines, S.A., Chappie, D.J., Lordo, R.A., Miller, B.D., Janke, R.J., Lindquist, H.A., Fox, K.R., Ernst, H.S. and Taft, S.C. (2014a) Assessment of relative potential for Legionella species or surrogates inhalation exposure from common water uses. Water Research 56, 203-213. Hines, S.A., Chappie, D.J., Lordo, R.A., Miller, B.D., Janke, R.J., Lindquist, H.A., Fox, K.R., Ernst, H.S. and Taft, S.C. (2014b) Assessment of relative potential for Legionella species or surrogates inhalation exposure from common water uses. Water Res. Hirneisen, K.A., Sharma, M. and Kniel, K.E. (2012) Human enteric pathogen internalization by root uptake into food crops. Foodborne pathogens and disease 9(5), 396-405. Hlavsa, M.C. and Brunkard, J.M. (2011) Surveillance for Waterborne Disease Outbreaks and Other Health Events Associated with Recreational Water--United States, 2007-2008, US Department of Health and Human Services, Centers for Disease Control and Prevention. Holmes, N.S. and Morawska, L. (2006) A review of dispersion modelling and its application to the dispersion of particles: an overview of different dispersion models available. Atmospheric Environment 40(30), 5902-5928. 375 Holvoet, K., De Keuckelaere, A., Sampers, I., Van Haute, S., Stals, A. and Uyttendaele, M. (2014) Quantitative study of cross-contamination with< i> Escherichia coli,< i> E. coli O157, MS2 phage and murine norovirus in a simulated fresh-cut lettuce wash process. Food Control 37, 218-227. Hood, B., Gardner, E., Barton, R., Gardiner, R., Beal, C., Hyde, R. and Walton, C. (2009) Decentralised development: the Ecovillage at Currumbin. Silva. Huang, W.-C., Chiou, C.-S., Chen, J.-H. and Shen, G.-H. (2010) Molecular epidemiology of Mycobacterium abscessus in a subtropcal chronic ventilatory setting. J Med Micro 59. Hugosson, A., Hjorth, M., Bernander, S., Claesson, B.E., Johansson, A., Larsson, H., Nolskog, P., Pap, J., Svensson, N. and Ulleryd, P. (2007) A community outbreak of Legionnaires’ disease from an industrial cooling tower: assessment of clinical features and diagnostic procedures. Scandinavian journal of infectious diseases 39(3), 217-224. Huston, R., Chan, Y.C., Gardner, T., Shaw, G. and Chapman, H. (2009) Characterisation of atmospheric deposition as a source of contaminants in urban rainwater tanks. Water Res 43(6), 1630-1640. Huttunen, K., Jussila, J., Hirvonen, M.R., Iivanainen, E. and Katila, M.L. (2001) Comparison of mycobacteria-induced cytotoxicity and inflammatory responses in human and mouse cell lines. Inhalation Toxicology 13(11), 977-991. Huttunen, K., Ruotsalainen, M., Iivanainen, E., Torkko, P., Katila, M.L. and Hirvonen, M.R. (2000) Inflammatory responses in RAW264.7 macrophages caused by mycobacteria isolated from moldy houses. Environmental Toxicology and Pharmacology 8(4), 237-244. Hyland, J., Hamlet, N., Saunders, C., Coppola, J. and Watt, J. (2008) Outbreak of Legionnaires’ disease in West Fife: Review of environmental guidelines needed. Public health 122(1), 79-83. Iakhiaeva, E., Howard, S.T., Brown Elliott, B.A., McNulty, S., Newman, K.L., Falkinham, J.O., 3rd, Williams, M., Kwait, R., Lande, L., Vasireddy, R., Turenne, C. and Wallace, R.J., Jr. (2016) Variable-number tandem-repeat analysis of respiratory and household water biofilm isolates of "Mycobacterium avium subsp. hominissuis" with establishment of a PCR database. J Clin Microbiol 54(4), 891-901. Iakhiaeva, E., McNulty, S., Brown Elliott, B.A., Falkinham, J.O., 3rd, Williams, M.D., Vasireddy, R., Wilson, R.W., Turenne, C. and Wallace, R.J., Jr. (2013) Mycobacterial interspersed repetitive-unit-variable-number tandem-repeat (MIRU-VNTR) genotyping of mycobacterium intracellulare for strain comparison with establishment of a PCR-based database. J Clin Microbiol 51(2), 409-416. 376 Ichijo, T., Izumi, Y., Nakamoto, S., Yamaguchi, N. and Nasu, M. (2014) Distribution and respiratory activity of mycobacteria in household water system of healthy volunteers in Japan. PloS one 9(10), e110554. Ichiyama, S., Shimokata, K. and Tsukamura, M. (1988) The isolation of Mycobacterium avium complex from soil, water, and dusts. Microbiology and immunology 32(7), 733-739. Inderlied, C., Kemper, C. and Bermudez, L. (1993) The Mycobacterium avium complex. Clinical Microbiology Reviews 6(3), 266. Inderlied, C., Kolonoski, P., Wu, M. and Young, L. (1989a) Amikacin, ciprofloxacin, and imipenem treatment for disseminated Mycobacterium avium complex infection of beige mice. Antimicrobial agents and chemotherapy 33(2), 176-180. Inderlied, C.B., Kownoski, P.T., Wu, M. and Young, L.S. (1989b) In vitro and in vivo activity of azithromycin (CP 62,993) against the Mycobacterium avium complex. Journal of Infectious Diseases 159(5), 994-997. ISO (1998) Water quality—detection and enumeration of Legionella. International standard 11731, 1998. ISO, Geneva, Switzerland. Isquith, I.R. and Winters, H. (1987) The microbial analysis of domestic cistern water in the US Virgin Islands. Desalination 66, 415-421. Ito, Y., Hirai, T., Fujita, K., Maekawa, K., Niimi, A., Ichiyama, S. and Mishima, M. (2015) Increasing patients with pulmonary Mycobacterium avium complex disease and associated underlying diseases in Japan. Journal of Infection and Chemotherapy 21(5), 352-356. Iwamoto, T., Nakajima, C., Nishiuchi, Y., Kato, T., Yoshida, S., Nakanishi, N., Tamaru, A., Tamura, Y., Suzuki, Y. and Nasu, M. (2012) Genetic diversity of Mycobacterium avium subsp. hominissuis strains isolated from humans, pigs, and human living environment. Infection, Genetics and Evolution 12(4), 846-852. Jackson, C.R. (2003) Changes in community properties during microbial succession. Oikos 101(2), 444-448. Jackson, C.R., Churchill, P.F. and Roden, E.E. (2001) Successional changes in bacterial assemblage structure during epilithic biofilm development. Ecology 82(2), 555-566. 377 Jahne, M.A., Rogers, S., Holsen, T.M., Grimberg, S.J. and Ramler, I. (2015) Emission and Dispersion of Bioaerosols from Dairy Manure Application Sites: Human Health Risk Assessment. Jahne, M.A., Rogers, S.W., Holsen, T.M. and Grimberg, S.J. (2014) Quantitative microbial risk assessment of bioaerosols from a manure application site. Aerobiologia 31(1), 73-87. Jenkins, P.A., Pattyn S.R. and Portaels, F. (1982) Diagnostic Bacteriology. In: The Biology of Mycobacteria, Ratledge and J. Stanford (Eds.), Academic Press, London. Jensen, E.S., Lundbye-Christensen, S., Pedersen, L., Sørensen, H.T. and Schønheyder, H.C. (2002) Seasonal variation in meningococcal disease in Denmark: relation to age and meningococcal phenotype. Scandinavian journal of infectious diseases 35(4), 226-229. Ji, B.H., Lounis, N., Truffotpernot, C. and Grosset, J. (1994) EFFECTIVENESS OF VARIOUS ANTIMICROBIAL AGENTS AGAINST MYCOBACTERIUM-AVIUM COMPLEX IN THE BEIGE MOUSE MODEL. Antimicrobial agents and chemotherapy 38(11), 2521-2529. Ji, B.H., Lounis, N., TruffotPernot, C. and Grosset, J. (1996) How effective is KRM-1648 in treatment of disseminated Mycobacterium avium complex infections in beige mice? Antimicrobial agents and chemotherapy 40(2), 437-442. Jiménez, B. and Asano, T. (2008) Water reuse: an international survey of current practice, issues and needs, IWA publishing. Jin, B.W., Saito, H. and Yoshii, Z. (1984) Environmental mycobacteria in Korea. Microbiology and immunology 28(6), 667-677. Jjemba, P.K., Johnson, W., Bukhari, Z. and LeChevallier, M.W. (2015) Occurrence and Control of Legionella in Recycled Water Systems. Pathogens 4(3), 470-502. Jjemba, P.K., Weinrich, L.A., Cheng, W., Giraldo, E. and LeChevallier, M.W. (2010) Regrowth of potential opportunistic pathogens and algae in reclaimed-water distribution systems. Applied and Environmental Microbiology 76(13), 4169-4178. Johnson, D., Lynch, R., Marshall, C., Mead, K. and Hirst, D. (2013) Aerosol Generation by Modern Flush Toilets. Aerosol Science and Technology 47(9), 1047-1057. Johnson, M.M. and Odell, J.A. (2014) Nontuberculous mycobacterial pulmonary infections. J Thorac Dis 6(3), 210-220. 378 Jones, A.Q., Dewey, C.E., Doré, K., Majowicz, S.E., McEwen, S.A., David, W.-T., Eric, M., Carr, D.J. and Henson, S.J. (2006) Public perceptions of drinking water: a postal survey of residents with private water supplies. BMC Public Health 6(1), 94. Jones, M.P. and Hunt, W.F. (2010) Performance of rainwater harvesting systems in the southeastern United States. Resources, Conservation and Recycling 54(10), 623-629. Jones, P., Rennison, L.M., Matthews, P., Collins, P. and Brown, A. (1981) The occurrence and significance to animal health of Leptospira, Mycobacterium, Escherichia coli, Brucella abortus and Bacillus anthracis in sewage and sewage sludges. Journal of hygiene 86(01), 129-137. Jones, R.J. and Jenkins, D.E. (1965) Mycobacteria isolated from soil. Canadian journal of microbiology 11(2), 127-133. Jordan, F.S., R; Riley, JJ; Yoklic, MR (2008) Effective removal of microbial contamination from harvested rainwater using a simple point of use filtration and UV-disinfection device. Urban Water Journal 5(3), 209-218. Jorgensen, J. (1977) Experimental infection with Mycobacterium avium, serotype 2, in pigs. 1. Intravenous inoculations. Acta Veterinaria Scandinavica 18(4), 532. Joseph, C. and Ricketts, C. (2008) Chapter 3: The Epidemiology of Legionnaires’ Disease. In: Heuner K and Swanson M. Legionella: Molecular microbiology, Caister Academic Press, Norfolk, UK. ISBN 978-1-904455-26-4 pp35-52. Joseph, C. and Ricketts, K. (2010) Legionnaires disease in Europe 2007-2008. Euro Surveill 15(8). Julien, E., Boobis, A.R., Olin, S.S. and Group, I.R.F.T.W. (2009) The key events dose-response framework: a cross-disciplinary mode-of-action based approach to examining dose-response and thresholds. Critical reviews in food science and nutrition 49(8), 682-689. Kaevska, M., Lvoncik, S., Slana, I., Kulich, P. and Kralik, P. (2014) Microscopy, Culture, and Quantitative Real-Time PCR Examination Confirm Internalization of Mycobacteria in Plants. Applied and Environmental Microbiology 80(13), 3888-3894. Kaevska, M., Slana, I., Kralik, P., Reischl, U., Orosova, J., Holcikova, A. and Pavlik, I. (2011) “Mycobacterium avium subsp. hominissuis” in neck lymph nodes of children and their environment examined by culture and triplex quantitative real-time PCR. Journal of clinical microbiology 49(1), 167-172. 379 Kahana, L.M., Kay, J.M., Yakrus, M.A. and Waserman, S. (1997) Mycobacterium avium complex infection in an immunocompetent young adult related to hot tub exposure. CHEST Journal 111(1), 242-245. Kamala, T., Paramasivan, C., Daniel Herbert, V.P. and Prabhakar, R. (1994a) Isolation and species level identification of mycobacteria in environmental samples from the BCG trial area of South India. Appl Environ Microbiology 60, 2180-2183. Kamala, T., Paramasivan, C., Herbert, D., Venkatesan, P. and Prabhakar, R. (1994b) Evaluation of procedures for isolation of nontuberculous mycobacteria from soil and water. Applied and Environmental Microbiology 60(3), 1021-1024. Kamala, T., Paramasivan, C.N., Herbert, D., Venkatesan, P. and Prabhakar, R. (1994c) Isolation and identification of environmental mycobacteria in the Mycobacterium bovis BCC trial area of South India. Applied and Environmental Microbiology 60(6), 2180-2183. Kansal, R.G., Gomez-Flores, R. and Mehta, R.T. (1998) Change in colony morphology influences the virulence as well as the biochemical properties of theMycobacterium aviumcomplex. Microbial pathogenesis 25(4), 203-214. Kanyok, T.P., Reddy, M.V., Chinnaswamy, J., Danziger, L.H. and Gangadharam, P.R.J. (1994) IN-VIVO ACTIVITY OF PAROMOMYCIN AGAINST SUSCEPTIBLE AND MULTIDRUG- RESISTANT MYCOBACTERIUM-TUBERCULOSIS AND MYCOBACTERIUM-AVIUM COMPLEX STRAINS. Antimicrobial agents and chemotherapy 38(2), 170-173. Karagiannis, I., Brandsema, P. and Van der Sande, M. (2009) Warm, wet weather associated with increased Legionnaires' disease incidence in The Netherlands. Epidemiol Infect 137(02), 181- 187. Kartalija, M., Ovrutsky, A.R., Bryan, C.L., Pott, G.B., Fantuzzi, G. and Thomas, J. (2013) Patients with nontuberculous mycobacterial lung disease exhibit unique body and immune phenotypes. Am J Respir Crit Care Med 187. Kasperbauer, S.H. and Daley, C.L. (2008) Diagnosis and treatment of infections due to Mycobacterium avium complex, p. 569. Katila, M., Iivanainen, E., Torkko, P., Kauppinen, J., Martikainen, P. and Väänänen, P. (1995) Isolation of potentially pathogenic mycobacteria in the Finnish environment. Scandinavian journal of infectious diseases. Supplementum 98, 9-11. Katz, S.M. and Hammel, J.M. (1987) The effect of drying, heat, and pH on the survival of Legionella pneumophila. Annals of Clinical & Laboratory Science 17(3), 150-156. 380 Kawamura, Y., Li, Y., Liu, H., Huang, X., Li, Z. and Ezaki, T. (2001) Bacterial population in Russian space station “Mir”. Microbiology and immunology 45(12), 819-828. Kazda, J. (1973) The importance of water for the spread of potentially pathogenic mycobacteria. I. Possibilities for the multiplication of mycobacteria. Zentralblatt fur Bakteriologie, Parasitenkunde, Infektionskrankheiten und Hygiene 158(2), 161-169. Keating, G.A. and McKone, T.E. (1993) Measurements and evaluation of the water-to-air transfer and air concentration for trichloroethylene in a shower chamber. ASTM SPECIAL TECHNICAL PUBLICATION 1205, 14-14. Kelly, D.G., Weir, R.D. and White, S.D. (2011) An investigation of roof runoff during rain events at the Royal Military College of Canada and potential discharge to Lake Ontario. Journal of Environmental Sciences 23(7), 1072-1078. Khan, K., Wang, J. and Marras, T.K. (2007) Nontuberculous mycobacterial sensitization in the United States - National trends over three decades. American Journal of Respiratory and Critical Care Medicine 176(3), 306-313. Khoor, A., Leslie, K.O., Tazelaar, H.D., Helmers, R.A. and Colby, T.V. (2001) Diffuse pulmonary disease caused by nontuberculous mycobacteria in immunocompetent people (hot tub lung). American Journal of Clinical Pathology 115(5), 755-762. Kilby, J.M., Gilligan, P.H., Yankaskas, J.R., Highsmith, W., Edwards, L.J. and Knowles, M.R. (1992) Nontuberculous mycobacteria in adult patients with cystic fibrosis. CHEST Journal 102(1), 70-75. Kim, R.D., Greenberg, D.E., Ehrmantraut, M.E., Guide, S.V., Ding, L. and Shea, Y. (2008) Pulmonary nontuberculous mycobacterial disease: prospective study of a distinct preexisting syndrome. Am J Respir Crit Care Med 178. Kim, R.H., Lee, S., Kim, Y.M. and Lee, J.H. (2004) Pollutants in rainwater runoff and their control by surface modification using TiO2. Journal of Industrial and Engineering Chemistry 10(1), 152-160. Kincaid, D., Solomon, K. and Oliphant, J. (1996) Drop size distributions for irrigation sprinklers. Transactions of the ASAE 39(3), 839-845. Kinzelman, J., Ng, C., Jackson, E., Gradus, S. and Bagley, R. (2003) Enterococci as indicators of Lake Michigan recreational water quality: comparison of two methodologies and their impacts on public health regulatory events. Appl Environ Microbiol 69(1), 92-96. 381 Kirschner Jr, R.A., Parker, B.C. and Falkinham III, J.O. (1992) Epidemiology of infection by nontuberculous mycobacteria: Mycobacterium avium, Mycobacterium intracellulare, and Mycobacterium scrofulaceum in acid, brown-water swamps of the Southeastern United States and their association with environmental variables. The American review of respiratory disease 145(2), 271-275. Kirschner, R.A., Jr., Parker, B.C. and Falkinham, J.O., 3rd (1992) Epidemiology of infection by nontuberculous mycobacteria. Mycobacterium avium, Mycobacterium intracellulare, and Mycobacterium scrofulaceum in acid, brown-water swamps of the southeastern United States and their association with environmental variables. Am Rev Respir Dis 145(2 Pt 1), 271-275. Kirschner, R.A., Parker, B.C. and Falkinham, J.O. (1999) Humic and fulvic acids stimulate the growth of Mycobacterium avium. FEMS microbiology ecology 30(4), 327-332. Klanicova, B., Seda, J., Slana, I., Slany, M. and Pavlik, I. (2013) The tracing of mycobacteria in drinking water supply systems by culture, conventional, and real time PCRs. Current microbiology 67(6), 725-731. Klanicova, B., Slana, I., Vondruskova, H., Kaevska, M. and Pavlik, I. (2011) Real-time quantitative PCR detection of Mycobacterium avium subspecies in meat products. Journal of Food Protection® 74(4), 636-640. Klanicova, B., Slany, M. and Slana, I. (2014) Analysis of sediments and plants from the system of five fishponds in the Czech Republic using culture and PCR methods. Science of The Total Environment 472, 851-854. Klausen, J., Giese, S., Fuursted, K. and Ahrens, P. (1997) Distribution of serotypes, IS901 and a 40 kDa protein in Mycobacterium avium complex strains isolated from man and animals in Denmark. Apmis 105(1‐6), 277-282. Kleeberg, H. and Nel, E. (1969) Porcine mycobacterial lymphadenitis. Jl S Afr Vet Med Assoc 40, 233-250. Kleeberg, H. and Nel, E.E. (1973) Occurrence of environmental atypical mycobacteria in South Africa. Ann. Soc. Belg. Med. Trop 53(4), 405-418. Kobayashi, M., Oana, K. and Kawakami, Y. (2014) Incidence of Legionella and heterotrophic bacteria in household rainwater tanks in Azumino, Nagano prefecture, Japan. Microbiology and Immunology 58(1), 15-21. 382 Koh, W.-J., Kwon, O.J., Jeon, K., Kim, T.S., Lee, K.S., Park, Y.K. and Bai, G.H. (2006) Clinical significance of nontuberculous mycobacteria isolated from respiratory specimens in Korea. CHEST Journal 129(2), 341-348. Koh, W.J., Jeong, B.H., Jeon, K., Lee, N.Y., Lee, K.S., Woo, S.Y., Shin, S.J. and Kwon, O.J. (2012) Clinical significance of the differentiation between Mycobacterium avium and Mycobacterium intracellulare in M. avium complex lung disease. Chest 142(6), 1482-1488. Kohl, R. (1974) Drop size distributions from medium-sized agricultural sprinklers. Transaction of the ASAE 17(4), 690-693. Koplan, J.D., DR; Swanston, WH; Tota, B (1978) Contaminated roof-collected rainwater as a possible cause of an outbreak of salmonellosis. J Hyg comb 81, 303. Koschel, D. (2006) Atypical mycobacteria as a cause of machine operator lung, hot tub lung and swimming pool alveolitis. Allergologie 29(11), 451-452. Krizova, K., Matlova, L., Horvathova, A., Moravkova, M., Beran, V., Boisselet, T., Babak, V., Slana, I. and Pavlik, I. (2010) Mycobacteria in the environment of pig farms in the Czech Republic between 2003 and 2007. Veterinarni Medicina 55(2), 55-69. Kuiper, M.W., Wullings, B.A., Akkermans, A.D., Beumer, R.R. and Van Der Kooij, D. (2004a) Intracellular proliferation of Legionella pneumophila in Hartmannella vermiformis in aquatic biofilms grown on plasticized polyvinyl chloride. Applied and Environmental Microbiology 70(11), 6826-6833. Kuiper, M.W., Wullings, B.A., Akkermans, A.D.L., Beumer, R.R. and Van Der Kooij, D. (2004b) Intracellular proliferation of Legionella pneumophila in Hartmannella vermiformis in aquatic biofilms grown on plasticized polyvinyl chloride. Appl Environ Microbiol 70(11), 6826- 6833. Kumpers, P., Tiede, A., Kirschner, P., Girke, J., Ganser, A. and Peest, D. (2008) Legionnaires' disease in immunocompromised patients: a case report of Legionella longbeachae pneumonia and review of the literature. J Med Microbiol 57(Pt 3), 384-387. Kunimoto, D., Peppler, M., Talbot, J., Phillips, P. and Shafran, S. (2003) Analysis of Mycobacterium avium complex isolates from blood samples of AIDS patients by pulsed-field gel electrophoresis. Journal of clinical microbiology 41(1), 498-499. Kura, F., Amemura-Maekawa, J., Yagita, K., Endo, T., Ikeno, M., Tsuji, H., Taguchi, M., Kobayashi, K., Ishii, E. and Watanabe, H. (2006) Outbreak of Legionnaires' disease on a cruise 383 ship linked to spa-bath filter stones contaminated with Legionella pneumophila serogroup 5. Epidemiol Infect 134(02), 385-391. Kus, B., Kandasamy, J., Vigneswaran, S. and Shon, H.K. (2010) Analysis of first flush to improve the water quality in rainwater tanks. Water Science and Technology 61(2), 421-428. Kwamanga, D., Swai, O., Agwanda, R. and Githui, W. (1995) Effect of non-tuberculous Mycobacteria infection on tuberculin results among primary school children in Kenya. East African medical journal 72(4), 222-227. Kyriakopoulos, A., Tassios, P., Matsiota-Bernard, P., Marinis, E., Tsaousidou, S. and Legakis, N. (1997) Characterization to species level of Mycobacterium avium complex strains from human immunodeficiency virus-positive and-negative patients. Journal of clinical microbiology 35(11), 3001-3003. Lai, K.K., Stottmeier, K.D., Sherman, I.H. and McCabe, W.R. (1984) Mycobacterial cervical lymphadenopathy: relation of etiologic agents to age. Jama 251(10), 1286-1288. Långmark, J., Storey, M.V., Ashbolt, N.J. and Stenström, T.A. (2005) Accumulation and fate of microorganisms and microspheres in biofilms formed in a pilot-scale water distribution system. Appl Environ Microbiol 71(2), 706-712. Lau, H. and Ashbolt, N. (2009a) The role of biofilms and protozoa in Legionella pathogenesis: implications for drinking water. J Appl Microbiol 107(2), 368-378. Lau, H.Y. and Ashbolt, N.J. (2009b) The role of biofilms and protozoa in Legionella pathogenesis: implications for drinking water. J Appl Microbiol 107(2), 368-378. Le Dantec, C., Duguet, J.-P., Montiel, A., Dumoutier, N., Dubrou, S. and Vincent, V. (2002a) Occurrence of mycobacteria in water treatment lines and in water distribution systems. Applied and Environmental Microbiology 68(11), 5318-5325. Le Dantec, C., Duguet, J.P., Montiel, A., Dumoutier, N., Dubrou, S. and Vincent, V. (2002b) Chlorine disinfection of atypical mycobacteria isolated from a water distribution system. Applied and Environmental Microbiology 68(3), 1025-1032. LeChevallier, M.W., Bukhari, Z., Jjemba, P., Johnson, W., Haas, C.N. and Hamilton, K.A. (2017) Development of a risk management strategy for Legionella in recycled water systems (WRF12- 05), WateReuse Research Foundation, Alexandria, VA. 384 Leckie, S.E. (2005) Methods of microbial community profiling and their application to forest soils. Forest Ecology and Management 220(1), 88-106. Ledwon, A., Szeleszczuk, P., Zwolska, Z., Augustynowicz-Kopec, E., Sapierzynski, R. and Kozak, M. (2008) Experimental infection of budgerigars (Melopsittacus undulatus) with five Mycobacterium species. Avian Pathology 37(1), 59-64. Lee, B.-Y., Chatterjee, D., Bozic, C., Brennan, P., Cohn, D., Bales, J., Harrison, S., Andron, L. and Orme, I. (1991) Prevalence of serum antibody to the type-specific glycopeptidolipid antigens of Mycobacterium avium in human immunodeficiency virus-positive and-negative individuals. Journal of clinical microbiology 29(5), 1026-1029. Lee, E.S., Lee, M.Y., Han, S.H. and Ka, J.O. (2008) Occurrence and molecular differentiation of environmental mycobacteria in surface waters. Journal of Microbiology and Biotechnology 18(7), 1207-1215. Lee, J.Y., Bak, G. and Han, M. (2012) Quality of roof-harvested rainwater--comparison of different roofing materials. Environ Pollut 162, 422-429. Lee, J.Y., Kim, H. and Han, M. (2016) Importance of maintenance in rainwater harvesting systems: a case study. Water Science and Technology: Water Supply 16(1), 97-103. Lee, J.Y., Yang, J.S., Han, M. and Choi, J. (2010) Comparison of the microbiological and chemical characterization of harvested rainwater and reservoir water as alternative water resources. Sci Total Environ 408(4), 896-905. Lee, M. and Visscher, J. (1992) Water harvesting- a guide for planners and project managers, IRC International Water and Sanitation Center, The Hague. Lehtola, M.J., Torvinen, E., Kusnetsov, J., Pitkänen, T., Maunula, L., Von Bonsdorff, C.H., Martikainen, P.J., Wilks, S.A., Keevil, C.W. and Miettinen, I.T. (2007) Survival of Mycobacterium avium, Legionella pneumophila, Escherichia coli, and caliciviruses in drinking water-associated biofilms grown under high-shear turbulent flow. Appl Environ Microbiol 73(9), 2854-2859. Lehtola, M.J., Torvinen, E., Miettinen, I.T. and Keevil, C.W. (2006) Fluorescence in situ hybridization using peptide nucleic acid probes for rapid detection of Mycobacterium avium subsp. avium and Mycobacterium avium subsp. paratuberculosis in potable-water biofilms. Applied and Environmental Microbiology 72(1), 848-853. Leoni, E., Legnani, P., Sabattini, M.B. and Righi, F. (2001) Prevalence of Legionella spp. in swimming pool environment. Water Res 35(15), 3749-3753. 385 Levesque, B., Pereg, D., Watkinson, E., Maguire, J.S., Bissonnette, L., Gingras, S., Rouja, P., Bergeron, M.G. and Dewailly, E. (2008) Assessment of microbiological quality of drinking water from household tanks in Bermuda. Can J Microbiol 54(6), 495-500. Levine, A.D. and Asano, T. (2004) Peer Reviewed: Recovering Sustainable Water from Wastewater. Environ Sci Technol 38(11), 201A-208A. Li, H., Chien, S.-H., Hsieh, M.-K., Dzombak, D.A. and Vidic, R.D. (2011) Escalating water demand for energy production and the potential for use of treated municipal wastewater. Environ Sci Technol 45(10), 4195-4200. Lighthart, B. (1973) Survival of airborne bacteria in a high urban concentration of carbon monoxide. Applied Microbiology 25(1), 86-91. Lighthart, B. (1989) A statistical model of laboratory death rate measurements for airborne bacteria. Aerobiologia 5(2), 138-144. Lighthart, B. (1994) Atmospheric Microbial Aerosols, pp. 285-303, Springer. Lighthart, B. and Frisch, A. (1976) Estimation of viable airborne microbes downwind from a point source. Appl Environ Microbiol 31(5), 700-704. Lighthart, B. and Kim, J. (1989) Simulation of airborne microbial droplet transport. Appl Environ Microbiol 55(9), 2349-2355. Lighthart, B. and Mohr, A. (1987) Estimating downwind concentrations of viable airborne microorganisms in dynamic atmospheric conditions. Appl Environ Microbiol 53(7), 1580-1583. Lighthart, B., Shaffer, B., Marthi, B. and Ganio, L. (1991) Trajectory of aerosol droplets from a sprayed bacterial suspension. Appl Environ Microbiol 57(4), 1006-1012. Lighthart, B. and Shaffer, B.T. (1997) Increased airborne bacterial survival as a function of particle content and size. Aerosol Science and Technology 27(3), 439-446. Lim, K.-Y., Hamilton, A.J. and Jiang, S.C. (2015) Assessment of public health risk associated with viral contamination in harvested urban stormwater for domestic applications. Science of The Total Environment 523, 95-108. Lim, K.-Y. and Jiang, S.C. (2013) Reevaluation of health risk benchmark for sustainable water practice through risk analysis of rooftop-harvested rainwater. Water Res 47(20). 386 Lim, Y., Kek, Y.R., Lim, P., Yap, H., Goh, T.V. and Ng, L. (2011) Environmental surveillance and molecular characterization of Legionella in tropical Singapore. Trop Biomed 28(1), 149-159. Lin, H., Xu, B., Chen, Y. and Wang, W. (2009) Legionella pollution in cooling tower water of air‐conditioning systems in Shanghai, China. J Appl Microbiol 106(2), 606-612. Lin, W., Yu, Z., Chen, X., Liu, R. and Zhang, H. (2012) Molecular characterization of natural biofilms from household taps with different materials: PVC, stainless steel, and cast iron in drinking water distribution system. Applied Microbiology and Biotechnology, 1-9. Lincoln, E.M. and Gilbert, L.A. (1972) Disease in Children due to Mycobacteria other than Mycobacterium tuberculosis American Review of Respiratory Disease 105(5), 683-714. Lindsay, D.S., Brown, A.W., Brown, D.J., Pravinkumar, S., Anderson, E. and Edwards, G.F. (2012) Legionella longbeachae serogroup 1 infections linked to potting compost. Journal of medical microbiology 61(2), 218-222. Lopes Leivas Leite, F. (2015) Understanding Mycobacterium avium subspecies hominissuis microaggregate mediated pathogenesis. Virulence 6(7), 675-676. Lopez-Varela, E., Garcia-Basteiro, A.L., Santiago, B., Wagner, D., van Ingen, J. and Kampmann, B. (2015) Non-tuberculous mycobacteria in children: muddying the waters of tuberculosis diagnosis. Lancet Respiratory Medicine 3(3), 244-256. Lopez, G.U., Gerba, C.P., Tamimi, A.H., Kitajima, M., Maxwell, S.L. and Rose, J.B. (2013) Transfer efficiency of bacteria and viruses from porous and nonporous fomites to fingers under different relative humidity conditions. Appl Environ Microbiol 79(18), 5728-5734. Lorian, V., Matuck, N. and Levy, J. (1975) Susceptibility of newborn and adult guinea pigs to Mycobacterium intracellulare and M. kansasii. Scandinavian journal of respiratory diseases 56(2), 103-108. Lounis, N., Ji, B.H. and TruffotPernot, C. (1997) Comparative activities of amikacin against Mycobacterium avium complex in nude and beige mice. Antimicrobial agents and chemotherapy 41(5), 1168-1169. Lu, J., Struewing, I., Vereen, E., Kirby, A., Levy, K., Moe, C. and Ashbolt, N. (2016) Molecular Detection of Legionella spp. and their associations with Mycobacterium spp., Pseudomonas aeruginosa and amoeba hosts in a drinking water distribution system. Journal of applied microbiology 120(2), 509-521. 387 Lu, J., Struewing, I., Yelton, S. and Ashbolt, N. (2015) Molecular survey of occurrence and quantity of Legionella spp., Mycobacterium spp., Pseudomonas aeruginosa and amoeba hosts in municipal drinking water storage tank sediments. Journal of applied microbiology 119(1), 278- 288. Lucas, M., Martínez, P. and Viedma, A. (2012) Experimental determination of drift loss from a cooling tower with different drift eliminators using the chemical balance method. international journal of refrigeration 35(6), 1779-1788. Lumb, R., Stapledon, R., Scroop, A., Bond, P., Cunliffe, D., Goodwin, A., Doyle, R. and Bastian, I. (2004) Investigation of spa pools associated with lung disorders caused by Mycobacterium avium complex in immunocompetent adults. Applied and Environmental Microbiology 70(8), 4906-4910. Lye, D.J. (1987) Bacterial levels in cistern water systems of northern Kentucky. JAWRA Journal of the American Water Resources Association 23(6), 1063-1068. Lye, D.J. (1992) Microbiology of rainwater cistern systems: A review. Journal of Environmental Science & Health Part A 27(8), 2123-2166. Lye, D.J. (2002) Health risks associated with consumption of untreated water from household roof catchment systems. Journal of the American Water Resources Association 38(5), 1301-1306. Lye, D.J. (2009) Rooftop runoff as a source of contamination: a review. Sci Total Environ 407(21), 5429-5434. Macomber, P.S. (2004) Guidelines on Rainwater Catchment Systems for Hawaii, University of Hawaii at Manoa, College of Tropical Agriculture & Human Resources, Manoa, Hawaii. Madigan, M.T., Martinko, J.M., Stahl, D.A. and Clark, D.P. (2012) Brock Biology of Microorganisms, Pearson Education Inc, San Francisco, CA. Madonia, P., Cangemi, M., Bellomo, S. and D'Alessandro, W. (2012) Influence of volcanic activity on the quality of water collected in roof water catchment systems at Stromboli Island (Italy). Journal of Geochemical Exploration. Maesaki, S., Kohno, S., Koga, H., Miyazaki, Y. and Kaku, M. (1993) A clinical comparison between Mycobacterium avium and Mycobacterium intracellulare infections. CHEST Journal 104(5), 1408-1411. 388 Makovcova, J., Slany, M., Babak, V., Slana, I. and Kralik, P. (2014) The water environment as a source of potentially pathogenic mycobacteria. Journal of water and health 12(2), 254-263. Mandarano, L. (2011) Sustainability in America's Cities. Economics, I.P.C.f.R. (ed), pp. 157-179. Mangione, E.J., Huitt, G., Lenaway, D., Beebe, J., Bailey, A., Figoski, M., Rau, M.P., Albrecht, K.D. and Yakrus, M.A. (2001) Nontuberculous mycobacterial disease following hot tub exposure. Emerging infectious diseases 7(6), 1039. Mansfield, K.G. and Lackner, A.A. (1997) Simian immunodeficiency virus–inoculated macaques acquire Mycobacterium avium from potable water during AIDS. Journal of Infectious Diseases 175(1), 184-187. Mansfield, K.G., Pauley, D., Young, H.L. and Lackner, A.A. (1995) Mycobacterium avium complex in macaques with AIDS is associated with a specific strain of simian immunodeficiency virus and prolonged survival after primary infection. Journal of Infectious Diseases 172(4), 1149- 1152. Marchetti, N., Criner, K. and Criner, G.J. (2004) Characterization of functional, radiologic and lung function recovery post-treatment of hot tub lung. A case report and review of the literature. Lung 182(5), 271-277. Marciano-Cabral, F., Jamerson, M. and Kaneshiro, E. (2010) Free-living amoebae,< i> Legionella and< i> Mycobacterium in tap water supplied by a municipal drinking water utility in the USA. Journal of water and health 8(1), 71-82. Marras, T.K., Wallace, R.J., Koth, L.L., Stulbarg, M.S., Cowl, C.T. and Daley, C.L. (2005) Hypersensitivity pneumonitis reaction to Mycobacterium avium in household water. CHEST Journal 127(2), 664-671. Marston, B.J., Lipman, H.B. and Breiman, R.F. (1994) Surveillance for Legionnaires' disease: risk factors for morbidity and mortality. Archives of Internal Medicine 154(21), 2417-2422. Martin-Casabona, N., Bahrmand, A., Bennedsen, J., Østergaard Thomsen, V., Curcio, M., Fauville-Dufaux, M., Feldman, K., Havelkova, M., Katila, M. and Köksalan, K. (2004) Non- tuberculous mycobacteria: patterns of isolation. A multi-country retrospective survey. The International Journal of Tuberculosis and Lung Disease 8(10), 1186-1193. Martinson, B. and Thomas, T. (2009) Quantifying the first-flush phenomenon: effects of first- flush on water yield and quality. 389 Masters, S. and Edwards, M. (2015) Increased lead in water associated with iron corrosion. Environmental Engineering Science 32(5), 361-369. Matlova, L., Dvorska, L., Ayele, W.Y., Bartos, M., Amemori, T. and Pavlik, I. (2005) Distribution of Mycobacterium avium complex isolates in tissue samples of pigs fed peat naturally contaminated with mycobacteria as a supplement. Journal of clinical microbiology 43(3), 1261-1268. Matlova, L., Dvorska, L., Bartl, J., Bartos, M., Ayele, W., Alexa, M. and Pavlik, I. (2003) Mycobacteria isolated from the environment of pig farms in the Czech Republic during the years 1996 to 2002. Veterinarni Medicina-UZPI (Czech Republic) 48, 343-357. Matsuyama, M., Ishii, Y., Yageta, Y., Ohtsuka, S., Ano, S., Matsuno, Y., Morishima, Y., Yoh, K., Takahashi, S., Ogawa, K., Hogaboam, C.M. and Hizawa, N. (2014) Role of Th1/Th17 Balance Regulated by T-bet in a Mouse Model of Mycobacterium avium Complex Disease. Journal of Immunology 192(4), 1707-1717. Matthews, P., Brown, A. and Collins, P. (1979) The Use of Polyvalent Sera for the Serotyping of Mycobacteria within the Mycobacterium avium– M. intracellulare-M. scrofulaceum Complex. J Appl Bacteriol 46(3), 425-430. May, T., Dailloux, M., Amiel, C., Laurain, C., Canton, P. and Faou, A. (1995) Mycobacterial infections in AIDS patients; Lorraine, France. Journal of the European Academy of Dermatology and Venereology 4(1), 100-103. Mayer, P.W. and DeOreo, W.B. (1999) Residential end uses of water, American Water Works Association. McCormick, D., Thorn, S., Milne, D., Evans, C., Stevenson, J., Llano, M. and Donaghy, M. (2011) Public health response to an outbreak of Legionnaires' disease in Edinburgh, United Kingdom, June 2012. Euro surveillance: bulletin Européen sur les maladies transmissibles= European communicable disease bulletin 17(28), 905-913. McDonald, J.H. (2009) Handbook of biological statistics, Sparky House Publishing Baltimore, MD. McDonough, E.A., Metzgar, D., Hansen, C.J., Myers, C.A. and Russell, K.L. (2007) A cluster of Legionella-associated pneumonia cases in a population of military recruits. Journal of Clinical Microbiology 45(6), 2075-2077. 390 McGavin, M., Mallmann, V., Mallmann, W. and Morrill, C. (1977) Pathological changes in calves injected intradermally with Mycobacterium intracellulare serotype Davis and M. avium serotype 2. Veterinary Pathology Online 14(1), 56-66. McShane, P.J. and Glassroth, J. (2015) Pulmonary Disease Due to Nontuberculous Mycobacteria Current State and New Insights. Chest 148(6), 1517-1527. McSwiggan, D.A. and Collins, C.H. (1974) The isolation of M. kansasii and M. xenopi from water systems. Tubercle 55. Medema, G., Wullings, B., Roeleveld, P. and Van Der Kooij, D. (2004) Risk assessment of Legionella and enteric pathogens in sewage treatment works. Water supply 4(2), 125-132. Meera, V. and Ahammed, M. (2006) Water quality of rooftop rainwater harvesting systems: a review. Aqua 55, 257-268. Mehta, R.T. (1996) Liposome encapsulation of clofazimine reduces toxicity in vitro and in vivo and improves therapeutic efficacy in the beige mouse model of disseminated Mycobacterium avium -M.intracellulare complex infection. Antimicrobial agents and chemotherapy 40(8), 1893- 1902. Meissner, G. and Anz, W. (1977) Sources of Mycobacterium avium complex infection resulting in human diseases. The American review of respiratory disease 116(6), 1057-1064. Mekkour, M., Khalil, E., Driss, B., Tai, J., Cohen, N., De Microbiologie, D., Environnement, D. and Pasteur, I. (2013) Legionella pneumophila: an environmental organism and accidental pathogen. Int J Sci Technol 2(2), 187-196. Melidis, P., Akratos, C.S., Tsihrintzis, V.A. and Trikilidou, E. (2007) Characterization of rain and roof drainage water quality in Xanthi, Greece. Environ Monit Assess 127(1-3), 15-27. Mena, K.D. and Gerba, C.P. (2009) Reviews of Environmental Contamination and Toxicology Vol 201, pp. 71-115, Springer. Mendez, C., Afshar, B., Kimney, K., Barrett, M. and Kirisits, M.J. (2010) Effect of roof material on water quality for rainwater harvesting systems, Texas Water Development Board. Mendez, C.B., Klenzendorf, J.B., Afshar, B.R., Simmons, M.T., Barrett, M.E., Kinney, K.A. and Kirisits, M.J. (2011) The effect of roofing material on the quality of harvested rainwater. Water Res 45(5), 2049-2059. 391 Merritt, A., Miles, R. and Bates, J. (1999) An outbreak of Campylobacter enteritis on an island resort, north Queensland. Commun Dis Intell 23(8), 215-219. Metcalf & Eddy (2007) Water Reuse : Issues, Technologies, and Applications: Issues, Technologies, and Applications, Mcgraw-hill, Michigan, USA. Mills, C.C. (1972) Occurrence of Mycobacterium other than Mycobacterium tuberculosis in the oral cavity and in sputum. Applied microbiology 24(3), 307-310. Milton, A., Stirzaker, S., Trungove, M., Knuckey, D., Martin, N., Hastie, C., Pennington, K., Sloan-Gardner, T., Fitzsimmons, G. and Knope, K. (2012) Australia's notifiable disease status, 2010: annual report of the National Notifiable Diseases Surveillance System. Communicable diseases intelligence quarterly report 36(1), 1-69. Minomo, S., Tachibana, K., Tsuyuguchi, K., Akira, M., Kitaichi, M. and Suzuki, K. (2015) A unique case of hot tub lung worsening during the winter. Internal Medicine 54(5), 491-495. Mirsaeidi, M. and Sadikot, R.T. (2015) Gender susceptibility to mycobacterial infections in patients with non-CF bronchiectasis. Int J Mycobacteriology 4. Montecalvo, M., Forester, G., Tsang, A., du Moulin, G. and Wormser, G. (1994) Colonisation of potable water with Mycobacterium avium complex in homes of HIV-infected patients. The Lancet 343(8913), 1639. Montero, J., Tarjuelo, J. and Carrión, P. (2003) Sprinkler droplet size distribution measured with an optical spectropluviometer. Irrigation Science 22(2), 47-56. Moore, G., Hewitt, M., Stevenson, D., Walker, J.T. and Bennett, A.M. (2015) Aerosolization of Respirable Droplets from a Domestic Spa Pool and the Use of MS-2 Coliphage and Pseudomonas aeruginosa as Markers for Legionella pneumophila. Appl Environ Microbiol 81(2), 555-561. Moraga-McHaley, S.A., Landen, M., Krapfl, H. and Sewell, C.M. (2013) Hypersensitivity pneumonitis with Mycobacterium avium complex among spa workers. International journal of occupational and environmental health 19(1), 55-61. Morales-Pinzón, T., García-Serna, M. and Flórez-Calderón, M. (2015) Quality of rainwater harvesting in urban systems: case study in Colombia. Water Practice and Technology 10(3), 424- 431. 392 Moritz, M.M., Flemming, H.-C. and Wingender, J. (2010) Integration of Pseudomonas aeruginosa and Legionella pneumophila in drinking water biofilms grown on domestic plumbing materials. International journal of hygiene and environmental health 213(3), 190-197. Muder, R.R. and Victor, L.Y. (2002) Infection due to Legionella species other than L. pneumophila. Clinical infectious diseases 35(8), 990-998. Mura, M., Bull, T.J., Evans, H., Sidi-Boumedine, K., McMinn, L., Rhodes, G., Pickup, R. and Hermon-Taylor, J. (2006) Replication and long-term persistence of bovine and human strains of Mycobacterium avium subsp. paratuberculosis within Acanthamoeba polyphaga. Applied and Environmental Microbiology 72(1), 854-859. Murrell, W. and Stewart, B. (1983) Botulism in New South Wales, 1980-1981. The Medical Journal of Australia 1(1), 13. Nakamura, K., Yokomizo, Y., Okutomo, M., Nishimori, K., Yugi, H. and Shoya, S. (1984) Light and electron microscopic observations on granulomatous lesions in pigs dosed with< i> Mycobacterium intracellulare. Journal of Comparative Pathology 94(4), 509-519. Nazarian, E.J., Bopp, D.J., Saylors, A., Limberger, R.J. and Musser, K.A. (2008) Design and implementation of a protocol for the detection of Legionella in clinical and environmental samples. Diagnostic microbiology and infectious disease 62(2), 125-132. Neil, K. and Berkelman, R. (2008) Increasing incidence of legionellosis in the United States, 1990–2005: changing epidemiologic trends. Clinical infectious diseases 47(5), 591-599. Nel, E.E. (1981) Mycobacterium avium-intracellulare complex serovars isolated in South Africa from humans, swine, and the environment. Review of Infectious Diseases 3(5), 1013-1020. Ng, V., Tang, P. and Fisman, D.N. (2008a) Our evolving understanding of legionellosis epidemiology: learning to count. Clinical infectious diseases 47(5), 600-602. Ng, V., Tang, P., Jamieson, F., Drews, S.J., Brown, S., Low, D.E., Johnson, C.C. and Fisman, D.N. (2008b) Going with the flow: legionellosis risk in Toronto, Canada is strongly associated with local watershed hydrology. EcoHealth 5(4), 482-490. Nguyen, T.M.N., Ilef, D., Jarraud, S., Rouil, L., Campese, C., Che, D., Haeghebaert, S., Ganiayre, F., Marcel, F. and Etienne, J. (2006) A community-wide outbreak of legionnaires disease linked to industrial cooling towers—how far can contaminated aerosols spread? Journal of Infectious Diseases 193(1), 102-111. 393 NHMRC-NRMMC (2004) Australian drinking water guidelines, Natl. Health Med. Res. Counc. Nat. Resour. Mag. Minist. Counc., Canberra, Australia. Nicolay, N., Boland, M., Ward, M., Hickey, L., Collins, C., Lynch, M., McCarthy, M. and O'DONNELL, J. (2010) Investigation of Pontiac-like illness in office workers during an outbreak of Legionnaires' disease, 2008. Epidemiol Infect 138(11), 1667-1673. Nishiuchi, Y., Maekura, R., Kitada, S., Tamaru, A., Taguri, T., Kira, Y., Hiraga, T., Hirotani, A., Yoshimura, K., Miki, M. and Ito, M. (2007) The recovery of Mycobacterium avium- intracellulare complex (MAC) from the residential bathrooms of patients with pulmonary MAC. Clinical Infectious Diseases 45(3), 347-351. Nishiuchi, Y., Tamaru, A., Kitada, S., Taguri, T., Matsumoto, S., Tateishi, Y., Yoshimura, M., Ozeki, Y., Matsumura, N., Ogura, H. and Maekura, R. (2009) Mycobacterium avium complex organisms predominantly colonize in the bathtub inlets of patients' bathrooms. Japanese Journal of Infectious Diseases 62(3), 182-186. Norby, B., Fosgate, G.T., Manning, E.J., Collins, M.T. and Roussel, A.J. (2007) Environmental mycobacteria in soil and water on beef ranches: association between presence of cultivable mycobacteria and soil and water physicochemical characteristics. Veterinary Microbiology 124(1), 153-159. Norton, C.D. and LeChevallier, M.W. (2000a) A pilot study of bacteriological population changes through potable water treatment and distribution. Applied and Environmental Microbiology 66(1), 268-276. Norton, C.D. and LeChevallier, M.W. (2000b) A pilot study of bacteriological population changes through potable water treatment and distribution. Appl Environ Microbiol 66. Norton, C.D., LeChevallier, M.W. and Falkinham, J.O. (2004) Survival of Mycobacterium avium in a model distribution system. Water Res 38. NRMMC-EPHC-AHMC (2006) Australian Guidelines for Water Recycling: Managing Health and Environmental Risks (Phase I), Australia. Nwachuku, N. and Gerba, C.P. (2004) Reviews of environmental contamination and toxicology, pp. 93-131, Springer. NYCDEP (2011) NYC Green Infrastructure Plan: 2011 Update, http://www.nyc.gov/html/dep/pdf/green_infrastructure/gi_annual_report_2012.pdf, New York, NY. 394 Nygård, K., Werner-Johansen, Ø., Rønsen, S., Caugant, D.A., Simonsen, Ø., Kanestrøm, A., Ask, E., Ringstad, J., Ødegård, R. and Jensen, T. (2008) An Outbreak of Legionnaires Disease Caused by Long‐Distance Spread from an Industrial Air Scrubber in Sarpsborg, Norway. Clinical infectious diseases 46(1), 61-69. O'Brien, D.P., Currie, B.J. and Krause, V.L. (2000) Nontuberculous mycobacterial disease in northern Australia: a case series and review of the literature. Clinical infectious diseases 31(4), 958-967. O'Brien, R.J., Geiter, L.J. and Snider Jr, D.E. (1987) The epidemiology of nontuberculous mycobacterial diseases in the United States. Results from a national survey. The American review of respiratory disease 135(5), 1007-1014. O'Hogain, S., McCarton, L., McIntyre, N., Pender, J. and Reid, A. (2012) Physicochemical and microbiological quality of harvested rainwater from an agricultural installation in Ireland. Water and Environment Journal 26(1), 1-6. O'Loughlin, R.E., Kightlinger, L., Werpy, M.C., Brown, E., Stevens, V., Hepper, C., Keane, T., Benson, R.F., Fields, B.S. and Moore, M.R. (2007) Restaurant outbreak of Legionnaires' disease associated with a decorative fountain: an environmental and case-control study. Bmc Infectious Diseases 7(1), 1. O'Neil, R., Bennett, G. and Kapoor, C. (2006) High‐resolution computed tomography of ‘hot tub’pneumonitis. Australasian radiology 50(1), 52-54. O'Toole, J., Keywood, M., Sinclair, M. and Leder, K. (2009) Risk in the mist? Deriving data to quantify microbial health risks associated with aerosol generation by water-efficient devices during typical domestic water-using activities. O'Toole, J., Leder, K. and Sinclair, M. (2008a) A series of exposure experiments-Recycled water and alternative water sources: Part B-Microbial transfer efficiency during machine clothes washing and microbial survival turf-grass experiments, Cooperative Research Centre for Water Quality and Treatment. O'Toole, J., Leder, K. and Sinclair, M. (2008b) A series of exposure experiments–recycled water and alternative water sources. Part A. Aerosolsizing and endotoxin experiments. , CRC for Water Quality and Treatment, Adelaide, Australia. Ohno, A., Kato, N., Sakamoto, R., Kimura, S. and Yamaguchi, K. (2008) Temperature-dependent parasitic relationship between Legionella pneumophila and a free-living amoeba (Acanthamoeba castellanii). Appl Environ Microbiol 74(14), 4585-4588. 395 Olitzki, A., Olitzki, Z. and Davis, C. (1968) Intracerebral Infection of Mice with M. avium and Atypical Mycobacteria. Pathobiology 31(5), 287-306. Oliver, A., Maiz, L., Cantón, R., Escobar, H., Baquero, F. and Gómez-Mampaso, E. (2001) Nontuberculous mycobacteria in patients with cystic fibrosis. Clinical Infectious Diseases 32(9), 1298-1303. Oliver, J.D. (2005) The viable but nonculturable state in bacteria. J Microbiol 43(1), 93-100. Oloya, J., Opuda-Asibo, J., Kazwala, R., Demelash, A., Skjerve, E., Lund, A., Johansen, T. and Djonne, B. (2008) Mycobacteria causing human cervical lymphadenitis in pastoral communities in the Karamoja region of Uganda. Epidemiology and infection 136(05), 636-643. Ovrutsky, A.R., Chan, E.D., Kartalija, M., Bai, X., Jackson, M., Gibbs, S., Falkinham, J.O., Iseman, M.D., Reynolds, P.R. and McDonnell, G. (2013) Cooccurrence of free-living amoebae and nontuberculous mycobacteria in hospital water networks, and preferential growth of Mycobacterium avium in Acanthamoeba lenticulata. Applied and Environmental Microbiology 79(10), 3185-3192. Paez-Rubio, T., Ramarui, A., Sommer, J., Xin, H., Anderson, J. and Peccia, J. (2007) Emission rates and characterization of aerosols produced during the spreading of dewatered class B biosolids. Environ Sci Technol 41(10), 3537-3544. Pagnier, I., Merchat, M., Raoult, D. and La Scola, B. (2009) Emerging Mycobacteria spp. in cooling towers. Emerg Infect Dis 15(1), 121-122. Palla, A., Gnecco, I. and Lanza, L. (2011) Non-dimensional design parameters and performance assessment of rainwater harvesting systems. Journal of Hydrology 401(1), 65-76. Palmer, C.J., Bonilla, G.F., Roll, B., Paszko-Kolva, C., Sangermano, L.R. and Fujioka, R.S. (1995) Detection of Legionella species in reclaimed water and air with the EnviroAmp Legionella PCR kit and direct fluorescent antibody staining. Appl Environ Microbiol 61(2), 407-412. Palmore, T.N., Stock, F., White, M., Bordner, M., Michelin, A., Bennett, J.E., Murray, P.R. and Henderson, D.K. (2009) A cluster of cases of nosocomial legionnaires disease linked to a contaminated hospital decorative water fountain. Infection Control & Hospital Epidemiology 30(08), 764-768. Pandey, D.N., Gupta, A.K. and Anderson, D.M. (2003) Rainwater harvesting as an adaptation to climate change. Current science 85(1), 46-59. 396 Parashar, D., Chauhan, D., Sharma, V., Chauhan, A., Chauhan, S. and Katoch, V. (2004) Optimization of procedures for isolation of mycobacteria from soil and water samples obtained in northern India. Applied and Environmental Microbiology 70(6), 3751-3753. Parker, B.C., Ford, M.A., Gruft, H. and Falkinham III, J.O. (1983) Epidemiology of Infection by Nontuberculous Mycobacteria: IV. Preferential Aerosolization of Mycobacterium intracellulare from Natural Waters 1–4. American Review of Respiratory Disease 128(4), 652-656. Parthuisot, N., West, N., Lebaron, P. and Baudart, J. (2010) High diversity and abundance of Legionella spp. in a pristine river and impact of seasonal and anthropogenic effects. Appl Environ Microbiol 76(24), 8201-8210. Pastoris, M.C., Ciceroni, L., Monaco, R.L., Goldoni, P., Mentore, B., Flego, G., Cattani, L., Ciarrocchi, S., Pinto, A. and Visca, P. (1997) Molecular epidemiology of an outbreak of Legionnaires' disease associated with a cooling tower in Genova-Sestri Ponente, Italy. European Journal of Clinical Microbiology and Infectious Diseases 16(12), 883-892. Pavlik, I., Falkinham 3rd, J. and Kazda, J. (2009a) Chapter 6: The occurrence of pathogenic and potentially pathogenic mycobacteria in animals and the role of the environment in the spread of infection. In: The ecology of mycobacteria: Impact on animal's and human's health, pp. 199-282, Springer, New York. Pavlik, I., Falkinham 3rd, J. and Kazda, J. (2009b) Chapter 3: Potentially pathogenic mycobacteria. In: The ecology of mycobacteria: Impact on animal's and human's health, pp. 21- 80, Springer, New York. Pavlik, I., Falkinham 3rd, J. and Kazda, J. (2009c) Chapter 5: Environments providing favourable conditions for the multiplication and transmission of mycobacteria. In: The ecology of mycobacteria: Impact on animal's and human's health, pp. 89-198, Springer, New York. Pavlik, I., Falkinham III, J. and Kazda, J. (2009d) Chapter 7: Transmission of mycobacteria from the environment to susceptible hosts. In: The Ecology of Mycobacteria: Impact on Animal's and Human's Health, pp. 283-312, Springer, New York. Pearson, C., Corner, L. and Lepper, A. (1977) Tuberculin sensitivity of cattle inoculated with atypical mycobacteria isolated from cattle, feral pigs and trough water. Australian veterinary journal 53(2), 67-71. Pedrosa, J., Florido, M., Kunze, Z., Castro, A., Portaels, F., McFadden, J., Silva, M. and Appelberg, R. (1994) Characterization of the virulence of Mycobacterium avium complex (MAC) isolates in mice. Clinical & Experimental Immunology 98(2), 210-216. 397 Pelletier, P., Du Moulin, G. and Stottmeier, K. (1988) Mycobacteria in public water supplies: comparative resistance to chlorine. Microbiological sciences 5(5), 147-148. Pepper, I.L., Brooks, J.P., Sinclair, R.G., Gurian, P.L. and Gerba, C.P. (2010) Pathogens and indicators in United States Class B biosolids: National and historic distributions. Journal of Environmental Quality 39(6), 2185-2190. Perronne, C. (1998) Mycobacterial infection models. Clinical Microbiology and Infection 4(10), 553-556. Peters, A.J., Weidner, K.L. and Howley, C.L. (2008) The chemical water quality in roof- harvested water cisterns in Bermuda. Journal of Water Supply Research and Technology-Aqua 57(3), 153-163. Peters, M., Müller, C., Rüsch-Gerdes, S., Seidel, C., Göbel, U., Pohle, H. and Ruf, B. (1995) Isolation of atypical mycobacteria from tap water in hospitals and homes: is this a possible source of disseminated MAC infection in AIDS patients? Journal of Infection 31(1), 39-44. Peterson, E.W. and Lighthart, B. (1977) Estimation of downwind viable airborne microbes from a wet cooling tower—Including settling. Microbial Ecology 4(1), 67-79. Pham, R.V., Vydareny, K.H. and Gal, A.A. (2003) High-resolution computed tomography appearance of pulmonary Mycobacterium avium complex infection after exposure to hot tub: case of hot-tub lung. Journal of thoracic imaging 18(1), 48-52. Phin, N., Parry-Ford, F., Harrison, T., Stagg, H.R., Zhang, N., Kumar, K., Lortholary, O., Zumla, A. and Abubakar, I. (2014) Epidemiology and clinical management of Legionnaires' disease. The Lancet Infectious Diseases 14(10), 1011-1021. Pickup, R., Rhodes, G., Arnott, S., Sidi-Boumedine, K., Bull, T., Weightman, A., Hurley, M. and Hermon-Taylor, J. (2005) Mycobacterium avium subsp. paratuberculosis in the catchment area and water of the River Taff in South Wales, United Kingdom, and its potential relationship to clustering of Crohn's disease cases in the city of Cardiff. Applied and Environmental Microbiology 71(4), 2130-2139. Pickup, R., Rhodes, G., Bull, T., Arnott, S., Sidi-Boumedine, K., Hurley, M. and Hermon-Taylor, J. (2006) Mycobacterium avium subsp. paratuberculosis in lake catchments, in river water abstracted for domestic use, and in effluent from domestic sewage treatment works: diverse opportunities for environmental cycling and human exposure. Applied and Environmental Microbiology 72(6), 4067-4077. 398 Pierce, E.S. (2009) Possible transmission of Mycobacterium avium subspecies paratuberculosis through potable water: lessons from an urban cluster of Crohn's disease. Gut pathogens 1(1), 1-5. Piso, R., Caruso, A. and Nebiker, M. (2007) Hose as a source of Legionella pneumonia. A new risk factor for gardeners? Journal of Hospital Infection 67(4), 396-397. Portaels, F. (1995) Epidemiology of mycobacterial diseases. Clinics in dermatology 13(3), 207. Poster, D.L. and Baker, J.E. (1996) Influence of submicron particles on hydrophobic organic contaminants in precipitation .1. Concentrations and distributions of polycyclic aromatic hydrocarbons and polychlorinated biphenyls in rainwater. Environ Sci Technol 30(1), 341-348. Powell, B. and Steadham, J.E. (1981) Improved technique for isolation of Mycobacterium kansasii from water. Journal of clinical microbiology 13(5), 969-975. Pravinkumar, S.E., G; Lindsay, D; Redmond, S; Stirling, J; House, R; Kerr, J; Anderson, E; Breen, D; Blatchford, O; McDonald, E; Brown, A (2010) A cluster of Legionnaires' disease caused by Legionella longbeachae linked to potting compost in Scotland, 2008-2009. Euro Surveillance 15(8). Prevots, D.R., Shaw, P.A., Strickland, D., Jackson, L.A., Raebel, M.A. and Blosky, M.A. (2010) Nontuberculous mycobacterial lung disease prevalence at four integrated health care delivery systems. Am J Respir Crit Care Med 182. Primm, T.P., Lucero, C.A. and Falkinham, J.O. (2004) Health impacts of environmental mycobacteria. Clinical Microbiology Reviews 17(1), 98-106. Prince, D.S., Peterson, D.D., Steiner, R.M., Gottlieb, J.E., Scott, R., Israel, H.L., Figueroa, W.G. and Fish, J.E. (1989) Infection with Mycobacterium avium complex in patients without predisposing conditions. New England Journal of Medicine 321(13), 863-868. Pruden, A., Edwards, M.A. and Falkinham, J.O.I. (2013) State of the Science and Research Needs for Opportunistic Pathogens in Premise Plumbing, Water Research Foundation, Denver, CO. Pryor, M., Springthorpe, S., Riffard, S., Brooks, T., Huo, Y., Davis, G. and Sattar, S. (2004a) Investigation of opportunistic pathogens in municipal drinking water under different supply and treatment regimes. Water Science and Technology 50(1), 83-90. Pryor, M., Springthorpe, S., Riffard, S., Brooks, T., Huo, Y., Davis, G. and Sattar, S. (2004b) Investigation of opportunistic pathogens in municipal drinking water under different supply and treatment regimes. Water Science & Technology 50(1), 83-90. 399 Rakić, A., Perić, J., Štambuk-Giljanović, N., Mikrut, A. and Bakavić, A.-S. (2011) Legionella species in year-round vs. seasonal accommodation water supply systems. Archives of Industrial Hygiene and Toxicology 62(4), 335-340. Rautiala, S., Torvinen, E., Torkko, P., Suomalainen, S., Nevalainen, A., Kalliokoski, P. and Katila, M.L. (2004) Potentially pathogenic, slow-growing mycobacteria released into workplace air during the remediation of buildings. Journal of Occupational and Environmental Hygiene 1(1), 1-6. Regli, S., Rose, J.B., Haas, C.N. and Gerba, C.P. (1991) Modeling the risk from Giardia and viruses in drinking water. Journal (American Water Works Association), 76-84. Reyneke, B., Dobrowsky, P., Ndlovu, T., Khan, S. and Khan, W. (2016) EMA-qPCR to monitor the efficiency of a closed-coupled solar pasteurization system in reducing Legionella contamination of roof-harvested rainwater. Science of The Total Environment 553, 662-670. Reznikov, M. and Dawson, D. (1971) Serological investigation of strains of Mycoacterium- intracellulare (Battey Bacillus) isolated from house-dusts. Med J Aust 1(13), 682-683. Reznikov, M. and Dawson, D. (1980) Mycobacteria of the intracellulare-scrofulaceum group in soils from the Adelaide area. Pathology 12(4), 525-528. Reznikov, M. and Leggo, J.H. (1974) Examination of soil in the Brisbane area for organisms of the Mycobacterium avium-intracellulare-scrofulaceum complex. Pathology 6(3), 269-273. Reznikov, M., Leggo, J.H. and Dawson, D.J. (1971a) Investigation by seroagglutination of strains of the Mycobacterium intracellulare-M. scrofulaceum group from house dusts and sputum in Southeastern Queensland. Am Rev Respir Dis 104. Reznikov, M., Leggo, J.H. and Dawson, D.J. (1971b) Investigation by seroagglutination of strains of the Mycobacterium intracellulare-M. scrofulaceum group from house dusts and sputum in Southeastern Queensland. Am Rev Respir Dis 104(6), 951-953. Rhoads, W.J., Pruden, A. and Edwards, M.A. (2016) Survey of green building water systems reveals elevated water age and water quality concerns. Environmental Science: Water Research & Technology 2(1), 164-173. Rhodes, G., Henrys, P., Thomson, B.C. and Pickup, R.W. (2013) Mycobacterium avium subspecies paratuberculosis is widely distributed in British soils and waters: implications for animal and human health. Environ Microbiol 15(10), 2761-2774. 400 Rhodes, G., Richardson, H., Hermon-Taylor, J., Weightman, A., Higham, A. and Pickup, R. (2014) Mycobacterium avium subspecies paratuberculosis: human exposure through environmental and domestic aerosols. Pathogens 3(3), 577-595. Rice, G., Wright, J.M., Boutin, B., Swartout, J., Rodgers, P., Niemuth, N. and Broder, M. (2005) Estimating the frequency of tap-water exposures to Mycobacterium avium complex in the US population with advanced aids. Journal of Toxicology and Environmental Health-Part a-Current Issues 68(11-12), 1033-1047. Richards, C.L., Broadaway, S.C., Eggers, M.J., Doyle, J., Pyle, B.H., Camper, A.K. and Ford, T.E. (2015) Detection of pathogenic and non-pathogenic bacteria in drinking water and associated biofilms on the crow reservation, Montana, USA. Microbial Ecology, 1-12. Richmond, R.E., Rickabaugh, J., Huffman, J. and Epperly, N. (1987) Colorectal cancer mortality and incidence in Campbell County, Kentucky. South Med J 80(8), 953-957. Ricketts, K., Charlett, A., Gelb, D., Lane, C., Lee, J. and Joseph, C. (2009) Weather patterns and Legionnaires' disease: a meteorological study. Epidemiol Infect 137(07), 1003-1012. Ricketts, K.D. (2007) Legionnaires’ disease in Europe: 2005-2006. Eurosurveillance Monthly Release 12, 12. Rickman, O.B., Ryu, J.H., Fidler, M.E. and Kalra, S. (2002) Hypersensitivity pneumonitis associated with Mycobacterium avium complex and hot tub use, pp. 1233-1237, Elsevier. Rindi, L. and Garzelli, C. (2014) Genetic diversity and phylogeny of Mycobacterium avium. Infection, Genetics and Evolution 21, 375-383. Ristola, M., Arbeit, R.D., von Reyn, C.F. and Horsburgh, C.R., Jr. (2015) Isolation of Mycobacterium avium from potable water in homes and institutions of patients with HIV infection in Finland and the United States. Biomed Res Int 2015, 713845. Ristola, M.A., von Reyn, C.F., Arbeit, R.D., Soini, H., Lumio, J., Ranki, A., Bühler, S., Waddell, R., Tosteson, A.N. and Falkinham, J. (1999) High rates of disseminated infection due to non- tuberculous mycobacteria among AIDS patients in Finland. Journal of Infection 39(1), 61-67. Rivière, D., Szczebara, F.M., Berjeaud, J.-M., Frère, J. and Héchard, Y. (2006) Development of a real-time PCR assay for quantification of Acanthamoeba trophozoites and cysts. Journal of microbiological methods 64(1), 78-83. 401 Rodgers, M., Blackstone, B., Reyes, A. and Covert, T. (1999) Colonisation of point of use water filters by silver resistant non-tuberculous mycobacteria. Journal of clinical pathology 52(8), 629. Rodrigo, S., Sinclair, M., Forbes, A., Cunliffe, D. and Leder, K. (2011) Drinking rainwater: a double-blinded, randomized controlled study of water treatment filters and gastroenteritis incidence. American Journal of Public Health 101(5), 842. Romanus, V., Hallander, H., Wåhlén, P., Olinder-Nielsen, A., Magnusson, P. and Juhlin, I. (1995) Atypical mycobacteria in extrapulmonary disease among children. Incidence in Sweden from 1969 to 1990, related to changing BCG-vaccination coverage. Tubercle and Lung Disease 76(4), 300-310. Roser, D., Van den Akker, B., Boase, S., Haas, C., Ashbolt, N. and Rice, S. (2014) Pseudomonas aeruginosa dose response and bathing water infection. Epidemiology and infection 142(03), 449- 462. Roser, D., van Den Akker, B., Boase, S., Haas, C., Ashbolt, N. and Rice, S. (2015) Dose– response algorithms for water-borne Pseudomonas aeruginosa folliculitis. Epidemiol Infect 143(07), 1524-1537. Rouil, L., Gardenas, G. and Marcel, F. (2004) Évaluation de la dispersion atmosphérique d'aérosols potentiellement contaminés lors de l’épidémie de légionellose de la région de Lens. NUMÉRO SPÉCIAL CONSACRÉ À LA LÉGIONELLOSE 2004(36/37), 182-184. Rowe, D.B. (2011) Green roofs as a means of pollution abatement. Environ Pollut 159(8-9), 2100-2110. Rusin, P.A., Rose, J.B., Haas, C.N. and Gerba, C.P. (1997) Risk assessment of opportunistic bacterial pathogens in drinking water. Reviews of environmental contamination and toxicology 152, 57-83. Ryan, M., Hamilton, K., Hamilton, M. and Haas, C.N. (2014) Evaluating the potential for a Helicobacter pylori drinking water guideline. Risk Analysis 34(9), 1651-1662. Sabria, M., Alvarez, J., Dominguez, A., Pedrol, A., Sauca, G., Salleras, L., Lopez, A., Garcia‐ Nuñez, M., Parron, I. and Barrufet, M. (2006) A community outbreak of Legionnaires' disease: evidence of a cooling tower as the source. Clinical Microbiology and Infection 12(7), 642-647. Saitanu, K. (1977) Studies on mycobacteria isolated from animals, with special reference to the agglutination test. Acta Pathologica Microbiologica Scandinavica Section B Microbiology 85(5), 303-307. 402 Saito, H., Tomioka, H., Sato, K., Tasaka, H. and Dawson, D.J. (1990) Identification of various serovar strains of Mycobacterium avium complex by using DNA probes specific for Mycobacterium avium and Mycobacterium intracellulare. J Clin Microbiol 28(8), 1694-1697. Saito, H. and Tsukamura, M. (1976) Mycobacterium intracellulare from public bath water. Japanese journal of microbiology 20(6), 561-563. Sakamoto, R., Ohno, A., Nakahara, T., Satomura, K., Iwanaga, S., Kouyama, Y., Kura, F., Kato, N., Matsubayashi, K., Okumiya, K. and Yamaguchi, K. (2009) Legionella pneumophila in Rainwater on Roads. Emerging Infectious Diseases 15(8), 1295-1297. Salah, I.B., Cayrou, C., Raoult, D. and Drancourt, M. (2009) Mycobacterium marseillense sp. nov., Mycobacterium timonense sp. nov. and Mycobacterium bouchedurhonense sp. nov., members of the Mycobacterium avium complex. International journal of systematic and evolutionary microbiology 59(11), 2803-2808. Salah, I.B. and Drancourt, M. (2010) Surviving within the amoebal exocyst: the Mycobacterium avium complex paradigm. BMC microbiology 10(1), 99. Sales-Ortells, H., Fernandez-Cassi, X., Timoneda, N., Dürig, W., Girones, R. and Medema, G. (2014) Health Risks Derived from Consumption of Lettuces Irrigated with Tertiary Effluent Containing Norovirus. Food Research International 68, 70-77. Sales-Ortells, H. and Medema, G. (2012) Screening-level risk assessment of Coxiella burnetii (Q fever) transmission via aeration of drinking water. Environ Sci Technol 46(7), 4125-4133. Sales-Ortells, H. and Medema, G. (2014) Screening-Level Microbial Risk Assessment of Urban Water Locations: A Tool for Prioritization. Environ Sci Technol 48(16), 9780-9789. Sales-Ortells, H. and Medema, G. (2015) Microbial Health Risks Associated with Exposure to Stormwater in a Water Plaza. Water Res. Salgado, M., Alfaro, M., Salazar, F., Badilla, X., Troncoso, E., Zambrano, A., González, M., Mitchell, R. and Collins, M. (2015) Application of cattle slurry containing Mycobacterium avium subsp. paratuberculosis (MAP) to grassland soil and its effect on the relationship between MAP and free-living amoeba. Veterinary Microbiology 175(1), 26-34. Sasaki, T., Matsumoto, N., Nakao, H., Katoh, T., Fukuda, Y., Nakazato, M. and Okayama, A. (2008) An outbreak of Legionnaires’ disease associated with a circulating bathwater system at a public bathhouse. I: a clinical analysis. Journal of Infection and Chemotherapy 14(2), 117-122. 403 Sawyer, C.N.M., Perry L.; Parkin, Gene F. (2003) Chemistry for Environmental Engineering and Science 5th Edition, Tata McGraw Hill Publishing Company Limited. Sax, H., Bloemberg, G., Hasse, B., Sommerstein, R., Kohler, P., Achermann, Y., Rossle, M., Falk, V., Kuster, S.P., Bottger, E.C. and Weber, R. (2015) Prolonged Outbreak of Mycobacterium chimaera Infection After Open-Chest Heart Surgery. Clinical Infectious Diseases 61(1), 67-75. Sayyad, G., Afyuni, M., Mousavi, S.F., Abbaspour, K.C., Hajabbasi, M.A., Richards, B.K. and Schulin, R. (2009) Effects of Cadmium, Copper, Lead, and Zinc Contamination on Metal Accumulation by Safflower and Wheat. Soil & Sediment Contamination 18(2), 216-228. Sazakli, E., Alexopoulos, A. and Leotsinidis, M. (2007) Rainwater harvesting, quality assessment and utilization in Kefalonia Island, Greece. Water Res 41(9), 2039-2047. Schaefer, W. (1965) Serologic identification and classification of the atypical mycobacteria by their agglutination. The American review of respiratory disease 92(6), 85-93. Schaefer, W., Davis, C.L. and Cohn, M. (1970) Pathogenicity of transparent, opaque, and rough variants of Mycobacterium avium in chickens and mice. The American review of respiratory disease 102(4), 499. Schaefer, W.B. (1967) Type-Specificity of Atypical Mycobacteria in Agglutination and Antibody Absorption Tests 1, 2. American Review of Respiratory Disease 96(6), 1165-1168. Schaefer, W.B. (1968) Incidence of the Serotypes of Mycobacterium avium and Atypical Mycobacteria in Human and Animal Diseases 1, 2. American Review of Respiratory Disease 97(1), 18-23. Schets, F., Italiaander, R., van den Berg, H. and de Roda Husman, A. (2010) Rainwater harvesting: quality assessment and utilization in The Netherlands. Journal of Water and Health 8(2), 224. Schets, F.M., Schijven, J.F. and de Roda Husman, A.M. (2011) Exposure assessment for swimmers in bathing waters and swimming pools. Water Res 45(7), 2392-2400. Schlech III, W.F., Gorman, G.W., Payne, M.C. and Broome, C.V. (1985) Legionnaires' disease in the Caribbean: an outbreak associated with a resort hotel. Archives of Internal Medicine 145(11), 2076. 404 Schlesinger, R.B. (1985) Comparative deposition of inhaled aerosols in experimental animals and humans: a review. Journal of Toxicology and Environmental Health, Part A Current Issues 15(2), 197-214. Schoen, M.E. and Ashbolt, N.J. (2011) An in-premise model for Legionella exposure during showering events. Water Res 45(18), 5826-5836. Schoen, M.E. and Garland, J. (2015) Review of pathogen treatment reductions for onsite non- potable reuse of alternative source waters. Microbial Risk Analysis. Schoen, M.E., Soller, J.A. and Ashbolt, N.J. (2011) Evaluating the importance of faecal sources in human-impacted waters. Water Res 45(8), 2670-2680. Schoen, M.E., Xue, X., Hawkins, T.R. and Ashbolt, N.J. (2014) Comparative Human Health Risk Analysis of Coastal Community Water and Waste Service Options. Environ Sci Technol 48, 9728-9736. Scholz, M. (2004) Case study: design, operation, maintenance and water quality management of sustainable storm water ponds for roof runoff. Bioresource Technology 95(3), 269-279. Schriewer, A., Horn, H. and Helmreich, B. (2008) Time focused measurements of roof runoff quality. Corrosion Science 50(2), 384-391. Schröder, K.-H., Kazda, J., Müller, K. and Müller, H. (1992) Isolation of Mycobacterium simiae from the environment. Zentralblatt für Bakteriologie 277(4), 561-564. Seinfeld, J.H. (1986) Atmospheric chemistry and physics of air pollution, John Wiley & Sons. Selby, K.A., Puckorius, P.R. and Helm, K.R. (1996) The use of reclaimed water in electric power stations and other industrial facilities. Water, Air, and Soil Pollution 90(1-2), 183-193. September, S., Brozel, V. and Venter, S. (2004) Diversity of nontuberculoid mycobacterium species in biofilms of urban and semiurban drinking water distribution systems. Appl Environ Microbiol 70. Shachor, Y., Dunovets, V., Poreh, S. and Shoham, C. (1997) Implications of simultaneous tests for tuberculin and non-tuberculous mycobacteria antigen in the elderly. Israel journal of medical sciences 33(3), 170-174. 405 Shannon, K., Lee, D.-Y., Trevors, J. and Beaudette, L. (2007) Application of real-time quantitative PCR for the detection of selected bacterial pathogens during municipal wastewater treatment. Science of The Total Environment 382(1), 121-129. Shao, Y., Chen, C., Song, H., Li, G., Liu, Q., Li, Y., Zhu, L., Martinez, L. and Lu, W. (2015) The epidemiology and geographic distribution of nontuberculous mycobacteria clinical isolates from sputum samples in the eastern region of China. PLoS Negl Trop Dis 9(3), e0003623. Sharifi, S., Haghshenas, M.M., Deksissa, T., Green, P., Hare, W. and Massoudieh, A. (2014) Storm Water Pollution Source Identification in Washington, DC, Using Bayesian Chemical Mass Balance Modeling. Journal of Environmental Engineering 140(3), 04013015 (04013011 pp.). Shitaye, J., Matlova, L., Horvathova, A., Moravkova, M., Dvorska-Bartosova, L., Treml, F., Lamka, J. and Pavlik, I. (2008) Mycobacterium avium subsp. avium distribution studied in a naturally infected hen flock and in the environment by culture, serotyping and IS901 RFLP methods. Veterinary Microbiology 127(1), 155-164. Shuval, H., Lampert, Y. and Fattal, B. (1997) Development of a risk assessment approach for evaluating wastewater reuse standards for agriculture. Water Science and Technology 35(11), 15- 20. Sibille, I., Sime-Ngando, T., Mathieu, L. and Block, J. (1998) Protozoan bacterivory and Escherichia coli survival in drinking water distribution systems. Appl Environ Microbiol 64(1), 197-202. Simmons, G., Jury, S., Thornley, C., Harte, D., Mohiuddin, J. and Taylor, M. (2008) A Legionnaires' disease outbreak: a water blaster and roof-collected rainwater systems. Water Res 42(6-7), 1449-1458. Simmons, G. and Smith, J. (1997) Roof water probable source of Salmonella infections. New Zealand Public Health Report 4(1), 5. Sinclair, M., Roddick, F., Nguyen, T., O'Toole, J. and Leder, K. (2016) Measuring water ingestion from spray exposures. Water Res 99, 1-6. Singer, E. and Rodda, G.M. (1965) Non-specific sensitization to old tuberculin: mycobacteria in water. Tubercle 46(2), 209-213. Skaza, A.T., Beskovnik, L., Storman, A., Ursic, S., Groboljsek, B. and Kese, D. (2010) Outbreak of Legionnaires’ disease in a nursing home, Slovenia, August 2010: preliminary report. Euro Surveill 15. 406 Slana, I., Kaevska, M., Kralik, P., Horvathova, A. and Pavlik, I. (2010) Distribution of Mycobacterium avium subsp. avium and M. a. hominissuis in artificially infected pigs studied by culture and IS901 and IS1245 quantitative real time PCR. Veterinary Microbiology 144(3-4), 437-443. Slana, I., Kralik, P., Kralova, A. and Pavlik, I. (2008) On-farm spread of Mycobacterium avium subsp. paratuberculosis in raw milk studied by IS900 and F57 competitive real time quantitative PCR and culture examination. International Journal of Food Microbiology 128(2), 250-257. Slifko, T.R., Huffman, D.E. and Rose, J.B. (1999) A most-probable-number assay for enumeration of infectious Cryptosporidium parvum oocysts. Applied and Environmental Microbiology 65(9), 3936-3941. Smithwick, R. (1976) Laboratory Manual for Acid-Fast Microscopy, Department of Health, Education, and Welfare, Public Health Service, Center for Disease Control, Atlanta, GA. Solomon, E.B., Yaron, S. and Matthews, K.R. (2002) Transmission of Escherichia coli O157: H7 from contaminated manure and irrigation water to lettuce plant tissue and its subsequent internalization. Appl Environ Microbiol 68(1), 397-400. Sood, A., Sreedhar, R., Kulkarni, P. and Nawoor, A.R. (2007) Hypersensitivity pneumonitis-like granulomatous lung disease with nontuberculous mycobacteria from exposure to hot water aerosols. Environmental health perspectives, 262-266. Spinks, A.T., Coombes, P., Dunstan, R. and Kuczera, G. (2003) Water quality treatment processes in domestic rainwater harvesting systems, pp. 227-234, Institution of Engineers, Australia, Wollongong, Australia. Spinks, A.T., Dunstan, R., Harrison, T., Coombes, P. and Kuczera, G. (2006) Thermal inactivation of water-borne pathogenic and indicator bacteria at sub-boiling temperatures. Water Res 40(6), 1326-1332. Ssematimba, A., Hagenaars, T.J. and De Jong, M.C. (2012) Modelling the wind-borne spread of highly pathogenic avian influenza virus between farms. PloS one 7(2), e31114. Stanford, J. and Paul, R. (1973) A preliminary report on some studies of environmental mycobacteria from Uganda. Ann Soc Belg Med Trop 53, 389-393. Steed, K.A. and Falkinham III, J.O. (2006) Effect of growth in biofilms on chlorine susceptibility of Mycobacterium avium and Mycobacterium intracellulare. Applied and Environmental Microbiology 72(6), 4007-4011. 407 Steele, T.M., CV; Sangster, N (1990) Distribution of Legionella longbeachae Serogroup 1 and other Legionellae in potting soils in Australia. Appl Environ Microbiol 56(10), 2984-2988. Steiner, M. and Boiler, M. (2006) Copper and zinc removal from roof runoff: from research to full-scale adsorber systems. Water Science and Technology 53(3), 199-207. Steinert, M., Birkness, K., White, E., Fields, B. and Quinn, F. (1998) Mycobacterium avium bacilli grow saprozoically in coculture with Acanthamoeba polyphaga and survive within cyst walls. Applied and Environmental Microbiology 64(6), 2256-2261. Stine, T.M., Harris, A.A., Levin, S., Rivera, N. and Kaplan, R.L. (1987) A pseudoepidemic due to atypical mycobacteria in a hospital water supply. Jama 258(6), 809-811. Storey, M., Ashbolt, J. and Stenström, T. (2004) Biofilms, thermophilic amoebae and Legionella pneumophila--a quantitative risk assessment for distributed water. Water science and technology: a journal of the International Association on Water Pollution Research 50(1), 77. Storey, M. and Ashbolt, N. (2003) A risk model for enteric virus accumulation and release from recycled water distribution pipe biofilms. Water supply 3(3), 93-100. Stormer, R.S. and Falkinham, J.O. (1989) Differences in antimicrobial susceptibility of pigmented and unpigmented colonial variants of Mycobacterium avium. Journal of clinical microbiology 27(11), 2459-2465. Strahl, E.D., Gillaspy, G.E. and Falkinham, I.I.I.J.O. (2001) Fluorescent Acid-Fast Microscopy for Measuring Phagocytosis of Mycobacterium avium, Mycobacterium intracellulare, and Mycobacterium scrofulaceum by Tetrahymena pyriformis and Their Intracellular Growth. Applied and Environmental Microbiology 67(10), 4432-4439. Struck, S. (2011) Rainwater harvesting for non-potable use and evidence of risk posed to human health, p. 17, British Columbia Centre for Disease Control, Vancouver, BC. Stumm, W. and Morgan, J.J. (1996) Aquatic Chemistry (3rd ed.), Wiley, New York. Stump, B., McBroom, M. and Darville, R. (2012) Demographics, practices and water quality from domestic potable rainwater harvesting systems. Journal of Water Supply: Research and Technology—AQUA 61(5), 261-271. Stypułkowska-Misiurewicz, H., Pancer, K., Krogulska, B. and Matuszewska, R. (2006) [Outbreak of hospital acquired Legionnaires' disease in patients of ophthalmic ward. Nosocomial Legionella infections for the first time observed in Poland]. Przeglad epidemiologiczny 61(4), 657-665. 408 Sugita, Y., Ishii, N., Katsuno, M., Yamada, R. and Nakajima, H. (2000) Familial cluster of cutaneous Mycobacterium avium infection resulting from use of a circulating, constantly heated bath water system. British Journal of Dermatology 142(4), 789-793. Tanaka, D., Niwatsukino, H., Oyama, T. and Nakajo, M. (2000) Progressing features of atypical mycobacterial infection in the lung on conventional and high resolution CT (HRCT) images. Radiation medicine 19(5), 237-245. Tanner, B.D., Brooks, J.P., Gerba, C.P., Haas, C.N., Josephson, K.L. and Pepper, I.L. (2008) Estimated occupational risk from bioaerosols generated during land application of class B biosolids. Journal of Environmental Quality 37(6), 2311-2321. Tateishi, Y., Hirayama, Y., Ozeki, Y., Nishiuchi, Y., Yoshimura, M., Kang, J., Shibata, A., Hirata, K., Kitada, S., Maekura, R., Ogura, H., Kobayashi, K. and Matsumoto, S. (2009) Virulence of Mycobacterium avium complex strains isolated from immunocompetent patients. Microbial pathogenesis 46(1), 6-12. Taylor, R.H., Falkinham III, J.O., Norton, C.D. and LeChevallier, M.W. (2000) Chlorine, chloramine, chlorine dioxide, and ozone susceptibility of Mycobacterium avium. Applied and Environmental Microbiology 66(4), 1702-1705. Teemusk, A. and Mander, Ü. (2007) Rainwater runoff quantity and quality performance from a greenroof: The effects of short-term events. Ecological Engineering 30(3), 271-277. Teltsch, B., Shuval, H.I. and Tadmor, J. (1980) Die-away kinetics of aerosolized bacteria from sprinkler application of wastewater. Appl Environ Microbiol 39(6), 1191-1197. Teng, J., Kumar, A., Gurian, P.L. and Olson, M.S. (2013) A Spreadsheet-Based Site Specific Risk Assessment Tool for Land-Applied Biosolids. Open Environmental Engineering Journal 6, 7-13. Tesh, M.J. and Miller, R.D. (1981) Amino acid requirements for Legionella pneumophila growth. Journal of Clinical Microbiology 13(5), 865-869. Thayer, W., Bozic, C.M., Camphausen, R.T. and McNEIL, M. (1990) Implications of antibodies to pyruvylated glucose in healthy populations for mycobacterioses and other infectious diseases. Journal of clinical microbiology 28(4), 714-718. Thomas, J.M. and Ashbolt, N.J. (2010) Do free-living amoebae in treated drinking water systems present an emerging health risk? Environmental science & technology 45(3), 860-869. 409 Thomas, R.B., Kirisits, M.J., Lye, D.J. and Kinney, K.A. (2014) Rainwater harvesting in the United States: a survey of common system practices. Journal of Cleaner Production (0). Thomas, T. (1998) Domestic water supply using rainwater harvesting. Building Research and Information 26(2), 94-101. Thomas, V. and McDonnell, G. (2007) Relationship between mycobacteria and amoebae: ecological and epidemiological concerns. Letters in Applied Microbiology 45(4), 349-357. Thomas, V., McDonnell, G., Denyer, S.P. and Maillard, J.-Y. (2010) Free-living amoebae and their intracellular pathogenic microorganisms: risks for water quality. FEMS Microbiology Reviews 34(3), 231-259. Thomson, R., Tolson, C., Carter, R., Coulter, C., Huygens, F. and Hargreaves, M. (2013a) Isolation of Nontuberculous Mycobacteria (NTM) from household water and shower aerosols in patients with pulmonary disease caused by NTM. Journal of clinical microbiology 51(9), 3006- 3011. Thomson, R.M. (2010) Changing epidemiology of pulmonary nontuberculous mycobacteria infections. Emerging infectious diseases 16(10), 1576. Thomson, R.M., Carter, R., Tolson, C., Coulter, C., Huygens, F. and Hargreaves, M. (2013b) Factors associated with the isolation of Nontuberculous mycobacteria (NTM) from a large municipal water system in Brisbane, Australia. BMC microbiology 13(1), 89. Thomson, R.M., Carter, R., Tolson, C., Coulter, C., Huygens, F. and Hargreaves, M. (2013c) Factors associated with the isolation of Nontuberculous mycobacteria (NTM) from a large municipal water system in Brisbane, Australia. BMC microbiology 13(1), 1. Thomson, R.M., Carter, R., Tolson, C., Coulter, C., Huygens, F. and Hargreaves, M. (2013d) Factors associated with the isolation of Nontuberculous mycobacteria (NTM) from a large municipal water system in Brisbane, Australia. BMC microbiology 13(1), 1-8. Thurman, R. (1995) Evaluation of rainwater stored in collection tanks. Australian Microbiologist 1(1), 20-22. Tichenor, W.S., Thurlow, J., McNulty, S., Brown-Elliott, B.A., Wallace Jr, R.J. and Falkinham III, J.O. (2012) Nontuberculous mycobacteria in household plumbing as possible cause of chronic rhinosinusitis. Emerg Infect Dis 18, 1612-1617. 410 Tobin-D’Angelo, M.J., Blass, M.A., del Rio, C., Halvosa, J.S., Blumberg, H.M. and Horsburgh, C.R. (2004a) Hospital water as a source of Mycobacterium avium complex isolates in respiratory specimens. Journal of Infectious Diseases 189(1), 98-104. Tobin-D’Angelo, M.J., Blass, M.A., del Rio, C., Halvosa, J.S., Blumberg, H.M. and Horsburgh, C.R. (2004b) Hospital water as a source of Mycobacterium avium complex isolates in respiratory specimens. J Inf Dis 189. Tomioka, H., Saito, H., Sato, K. and Dawson, D.J. (1993a) Comparison of the virulence for mice of Mycobacterium avium and Mycobacterium intracellulare identified by DNA probe test. Microbiology and immunology 37(4), 259-264. Tomioka, H., Sato, K., Saito, H. and Hidaka, T. (1993b) Therapeutic efficacy of a benzoxazinorifamycin, KRM-1648, combined with a immunopotentiator, LC9018, in Mycobacterium intracellulare infection induced in beige mice. Kekkaku : [Tuberculosis] 68(12), 751-754. Tomioka, H., Sato, K., Sano, C., Akaki, T., Shimizu, T., Kajitani, H. and Saito, H. (1997) Effector molecules of the host defence mechanism against Mycobacterium avium complex: The evidence showing that reactive oxygen intermediates, reactive nitrogen intermediates, and free fatty acids each alone are not decisive in expression of macrophage antimicrobial activity against the parasites. Clinical and Experimental Immunology 109(2), 248-254. Tortoli, E. (2003) Impact of genotypic studies on mycobacterial taxonomy: the new mycobacteria of the 1990s. Clinical Microbiology Reviews 16(2), 319-354. Tortoli, E., Rindi, L., Garcia, M.J., Chiaradonna, P., Dei, R., Garzelli, C., Kroppenstedt, R.M., Lari, N., Mattei, R. and Mariottini, A. (2004) Proposal to elevate the genetic variant MAC-A, included in the Mycobacterium avium complex, to species rank as Mycobacterium chimaera sp. nov. International journal of systematic and evolutionary microbiology 54(4), 1277-1285. Torvinen, E., Meklin, T., Torkko, P., Suomalainen, S., Reiman, M., Katila, M.-L., Paulin, L. and Nevalainen, A. (2006) Mycobacteria and fungi in moisture-damaged building materials. Applied and Environmental Microbiology 72(10), 6822-6824. Torvinen, E., Suomalainen, S., Lehtola, M.J., Miettinen, I.T., Zacheus, O., Paulin, L., Katila, M.- L. and Martikainen, P.J. (2004a) Mycobacteria in water and loose deposits of drinking water distribution systems in Finland. Appl Environ Microbiol 70. Torvinen, E., Suomalainen, S., Lehtola, M.J., Miettinen, I.T., Zacheus, O., Paulin, L., Katila, M.L. and Martikainen, P.J. (2004b) Mycobacteria in water and loose deposits of drinking water distribution systems in Finland. Applied and Environmental Microbiology 70(4), 1973-1981. 411 Torvinen, E., Suomalainen, S., Paulin, L. and Kusnetsov, J. (2014) Mycobacteria in Finnish cooling tower waters. Apmis 122(4), 353-358. Torvinen, E., Torkko, P. and Rintala, A.N. (2010) Real-time PCR detection of environmental mycobacteria in house dust. Journal of microbiological methods 82(1), 78-84. Travaline, J.M. and Kelsen, S.G. (2003) Hypersensitivity pneumonitis associated with hot tub use. Archives of internal medicine 163(18), 2250-2250. Travis, T.C., Brown, E.W., Peruski, L.F., Siludjai, D., Jorakate, P., Salika, P., Yang, G., Kozak, N.A., Kodani, M. and Warner, A.K. (2012) Survey of legionella species found in Thai soil. International journal of microbiology 2012. Tsang, A.Y., Denner, J.C., Brennan, P.J. and McClatchy, J.K. (1992) Clinical and epidemiological importance of typing of Mycobacterium avium complex isolates. Journal of clinical microbiology 30(2), 479-484. Tsukamura, M., Mizuno, S., Murata, H., Nemoto, H. and Yugi, H. (1974) A comparative study of mycobacteria from patients' room dusts and from sputa of tuberculous patients. Japanese journal of microbiology 18(4), 271-277. Tsukamura, M., Mizuno, S. and Toyama, H. (1984) [Mycobacteria from dusts of Japanese houses]. Kekkaku:[Tuberculosis] 59(12), 625-631. Tuffley, R., Leggo, J.H., Simmons, G. and Tammemägi, L. (1973) Studies on the virulence of Mycobacterium intracellulare serotype VI for pigs. Journal of Comparative Pathology 83(4), 467-471. Tuffley, R.H., JD (1980) Isolation of Mycobacterium avium-M. intracellulare-M.scrofulaceum complex from tank water in Queensland, Australia. Appl Environ Microbiol 39(1), 48-53. Turenne, C.Y., Wallace, R. and Behr, M.A. (2007) Mycobacterium avium in the postgenomic era. Clinical Microbiology Reviews 20(2), 205-229. TWDB (2005) The Texas Manual on Rainwater Harvesting, The Texas Water Development Board, Austin, Texas. Urabe, K. and Saito, H. (1962) 'Atypical’mycobacteria from the water of public baths. Japanese Journal of Bacteriology 17, 589-590. 412 US Census (2007) Vintage 2007: National Tables. USEPA (1982) Estimating microorganism densities in aerosols from spray irrigation of wastewater, US Environmental Protection Agency. USEPA (1984) National Statistical Assessment of Rural Water Conditions, Cornell University, Department of Rural Sociology. USEPA (2009) Drinking Water Contaminant Candidate List 3-Final, Federal Register. USEPA (2011a) Exposure Factors Handbook, Washington, DC. USEPA (2011b) Recreational Water Criteria. USEPA (2013) Rainwater Harvesting: Conservation, Credit, Codes, and Cost Literature Review and Case Studies, Washington, DC. USEPA (2015) Drinking Water Contaminant Candidate List 4- Draft, pp. 6076-6084, Federal Register. Vaerewijck, M.J.M., Huys, G., Palomino, J.C., Swings, J. and Portaels, F. (2005) Mycobacteria in drinking water distribution systems: ecology and significance for human health. FEMS Microbiology Reviews 29(5), 911-934. Valster, R.M., Wullings, B.A., van den Berg, R. and van der Kooij, D. (2011) Relationships between free-living protozoa, cultivable Legionella spp., and water quality characteristics in three drinking water supplies in the Caribbean. Appl Environ Microbiol 77(20), 7321-7328. van der Sterren, M., Rahman, A. and Dennis, G. (2013) Quality and Quantity Monitoring of Five Rainwater Tanks in Western Sydney, Australia. Journal of Environmental Engineering 139(3), 332-340. van Dijk, A.I., Beck, H.E., Crosbie, R.S., Jeu, R.A., Liu, Y.Y., Podger, G.M., Timbal, B. and Viney, N.R. (2013) The Millennium Drought in southeast Australia (2001–2009): Natural and human causes and implications for water resources, ecosystems, economy, and society. Water Resources Research 49(2), 1040-1057. Van Ingen, J., Boeree, M., Dekhuijzen, P. and Van Soolingen, D. (2009) Environmental sources of rapid growing nontuberculous mycobacteria causing disease in humans. Clin Micro Inf 15. 413 Van Leuken, J., Swart, A., Havelaar, A., Van Pul, A., Van der Hoek, W. and Heederik, D. (2015) Atmospheric dispersion modelling of bioaerosols that are pathogenic to humans and livestock–A review to inform risk assessment studies. Microbial Risk Analysis. Van Oss, C., Gillman, C. and A, N. (1975) Phagocytic Engulfment and Cell Adhesivness, Marcel Dekker, Inc., New York. Vanaclocha, H., Guiral, S., Morera, V., Calatayud, M., Castellanos, M., Moya, V., Jerez, G. and Gonzalez, F. (2012) Preliminary report: outbreak of Legionnaires disease in a hotel in Calp, Spain, update on 22 February 2012. Euro Surveill 17(8), 31-33. Vansnick, E., de Rijk, P., Vercammen, F., Geysen, D., Rigouts, L. and Portaels, F. (2004) Newly developed primers for the detection of Mycobacterium avium subspecies paratuberculosis. Veterinary Microbiology 100(3), 197-204. Vialle, C., Sablayrolles, C., Lovera, M., Huau, M.-C., Jacob, S. and Montrejaud-Vignoles, M. (2012) Water quality monitoring and hydraulic evaluation of a household roof runoff harvesting system in France. Water resources management, 1-9. Vialle, C., Sablayrolles, C., Lovera, M., Jacob, S., Huau, M.C. and Montrejaud-Vignoles, M. (2011) Monitoring of water quality from roof runoff: Interpretation using multivariate analysis. Water Res 45(12), 3765-3775. Viallier, J. and Viallier, G. (1973) Inventaire des mycobacteries de la nature. Ann Soc Belge Med Trop 4, 361-371. Viau, E., Bibby, K., Paez-Rubio, T. and Peccia, J. (2011) Toward a consensus view on the infectious risks associated with land application of sewage sludge. Environ Sci Technol 45(13), 5459-5469. Viau, E. and Peccia, J. (2009) Survey of wastewater indicators and human pathogen genomes in biosolids produced by class A and class B stabilization treatments. Appl Environ Microbiol 75(1), 164-174. Victor, L.Y., Plouffe, J.F., Pastoris, M.C., Stout, J.E., Schousboe, M., Widmer, A., Summersgill, J., File, T., Heath, C.M. and Paterson, D.L. (2002) Distribution of Legionella species and serogroups isolated by culture in patients with sporadic community-acquired legionellosis: an international collaborative survey. Journal of Infectious Diseases 186(1), 127-128. Villarreal, E.L. and Dixon, A. (2005) Analysis of a rainwater collection system for domestic water supply in Ringdansen, Norrkoping, Sweden. Building and Environment 40(9), 1174-1184. 414 Von Baum, H., Härter, G., Essig, A., Lück, C., Gonser, T., Embacher, A. and Brockmann, S. (2010) Preliminary report: outbreak of Legionnaires disease in the cities of Ulm and Neu-Ulm in Germany, December 2009–January 2010. Euro Surveill 15(4), 19472. von Reyn, C., Arbeit, R., Horsburgh, C., Ristola, M., Waddell, R., Tvaroha, S., Samore, M., Hirschhorn, L., Lumio, J. and Lein, A. (2002) Sources of disseminated Mycobacterium avium infection in AIDS. Journal of Infection 44(3), 166-170. von Reyn C, F., Horsburgh, C., Olivier, K., Barnes, P., Waddell, R., Warren, C., Tvaroha, S., Jaeger, A., Lein, A. and Alexander, L. (2001) Skin test reactions to Mycobacterium tuberculosis purified protein derivative and Mycobacterium avium sensitin among health care workers and medical students in the United States. The International Journal of Tuberculosis and Lung Disease 5(12), 1122-1128. von Reyn, C.F., Barber, T.W., Arbeit, R.D., Sox, C.H., O'Connor, G.T., Brindle, R.J., Gilks, C.F., Hakkarainen, K., Ranki, A. and Bartholomew, C. (1993a) Evidence of previous infection with Mycobacterium avium-Mycobacterium intracellulare complex among healthy subjects: an international study of dominant mycobacterial skin test reactions. Journal of Infectious Diseases 168(6), 1553-1558. von Reyn, C.F., Marlow, J., Arbeit, R., Barber, T. and Falkinham, J. (1994) Persistent colonisation of potable water as a source of Mycobacterium avium infection in AIDS. The Lancet 343(8906), 1137-1141. von Reyn, C.F., Waddell, R.D., Eaton, T., Arbeit, R.D., Maslow, J.N., Barber, T.W., Brindle, R.J., Gilks, C.F., Lumio, J. and Lähdevirta, J. (1993b) Isolation of Mycobacterium avium complex from water in the United States, Finland, Zaire, and Kenya. Journal of clinical microbiology 31(12), 3227-3230. Vugia, D.J., Jang, Y., Zizek, C., Ely, J., Winthrop, K.L. and Desmond, E. (2005) Mycobacteria in nail salon whirlpool footbaths, California. Emerging infectious diseases 11(4), 616. Waddell, L., Rajić, A., STÄRK, K. and McEWEN, S. (2015) The zoonotic potential of Mycobacterium avium ssp. paratuberculosis: a systematic review and meta-analyses of the evidence. Epidemiology and infection, 1-23. Waggie, K.S., Wagner, J.E. and Lentsch, R.H. (1983) EXPERIMENTAL MURINE INFECTIONS WITH A MYCOBACTERIUM-AVIUM-INTRACELLULARE COMPLEX ORGANISM ISOLATED FROM MICE. Laboratory Animal Science 33(3), 254-257. Wallace Jr, R.J., Zhang, Y., Brown, B.A., Dawson, D., Murphy, D.T., Wilson, R. and Griffith, D.E. (1998) Polyclonal Mycobacterium avium complex infections in patients with nodular bronchiectasis. American Journal of Respiratory and Critical Care Medicine 158(4), 1235-1244. 415 Wallace, R.J., Iakhiaeva, E., Williams, M.D., Brown-Elliott, B.A., Vasireddy, S., Vasireddy, R., Lande, L., Peterson, D.D., Sawicki, J. and Kwait, R. (2013) Absence of Mycobacterium intracellulare and presence of Mycobacterium chimaera in household water and biofilm samples of patients in the United States with Mycobacterium avium complex respiratory disease. Journal of clinical microbiology 51(6), 1747-1752. Waller, D. and Inman, D. (1982) Rain Water as an Alternative Source in Nova Scotia, pp. 202- 209, Honolulu, Hawaii. Walser, S.M., Gerstner, D.G., Brenner, B., Höller, C., Liebl, B. and Herr, C.E. (2014) Assessing the environmental health relevance of cooling towers–a systematic review of legionellosis outbreaks. Int J Hyg Environ Health 217(2), 145-154. Wang, H., Edwards, M., Falkinham, J.O. and Pruden, A. (2012a) Molecular survey of the occurrence of Legionella spp., Mycobacterium spp., Pseudomonas aeruginosa, and amoeba hosts in two chloraminated drinking water distribution systems. Applied and Environmental Microbiology 78(17), 6285-6294. Wang, H., Edwards, M.A., Falkinham III, J.O. and Pruden, A. (2013) Probiotic approach to pathogen control in premise plumbing systems? A review. Environmental science & technology 47(18), 10117-10128. Wang, H., Masters, S., Edwards, M.A., Falkinham, J.O. and Pruden, A. (2014a) Effect of Disinfectant, Water Age, and Pipe Materials on Bacterial and Eukaryotic Community Structure in Drinking Water Biofilm. Environ Sci Technol 48(3), 1426-1435. Wang, H., Masters, S., Edwards, M.A., Falkinham, J.O. and Pruden, A. (2014b) Effect of Disinfectant, Water Age, and Pipe Materials on Bacterial and Eukaryotic Community Structure in Drinking Water Biofilm. Environmental science & technology. Wang, H., Masters, S., Hong, Y., Stallings, J., Falkinham, J.O., Edwards, M.A. and Pruden, A. (2012b) Effect of Disinfectant, Water Age, and Pipe Material on Occurrence and Persistence of Legionella, mycobacteria, Pseudomonas aeruginosa, and Two Amoebas. Environ Sci Technol 46(21), 11566-11574. Warren, W.J. and Miller, R.D. (1979) Growth of Legionnaires disease bacterium (Legionella pneumophila) in chemically defined medium. Journal of Clinical Microbiology 10(1), 50-55. Wayne, L., Good, R., Tsang, A., Butler, R., Dawson, D., Groothuis, D., Gross, W., Hawkins, J., Kilburn, J. and Kubin, M. (1993) Serovar determination and molecular taxonomic correlation in Mycobacterium avium, Mycobacterium intracellulare, and Mycobacterium scrofulaceum: a cooperative study of the International Working Group on Mycobacterial Taxonomy. International journal of systematic and evolutionary microbiology 43(3), 482-489. 416 Wei, J., Jin, Y., Sims, T. and Kniel, K.E. (2011) Internalization of murine norovirus 1 by Lactuca sativa during irrigation. Appl Environ Microbiol 77(7), 2508-2512. Weir, M.H. and Haas, C.N. (2009) Quantification of the Effects of Age on the Dose Response of Variola major in Suckling Mice. Human and Ecological Risk Assessment 15(6), 1245-1256. Weiss, C.H. and Glassroth, J. (2012) Pulmonary disease caused by nontuberculous mycobacteria. Expert Review of Respiratory Medicine 6(6), 597-613. Wells, K. (2013) Australian weather and seasons - a variety of climates, Australian Government. Wendt, S.L., George, K.L., Parker, B.C., Gruft, H. and Falkinham, J.O., 3rd (1980) Epidemiology of infection by nontuberculous Mycobacteria. III. Isolation of potentially pathogenic mycobacteria from aerosols. Am Rev Respir Dis 122(2), 259-263. Whan, L., Ball, H., Grant, I. and Rowe, M. (2005a) Development of an IMS–PCR assay for the detection of Mycobacterium avium ssp. paratuberculosis in water. Letters in Applied Microbiology 40(4), 269-273. Whan, L., Ball, H.J., Grant, I.R. and Rowe, M.T. (2005b) Occurrence of Mycobacterium avium subsp. paratuberculosis in untreated water in Northern Ireland. Applied and Environmental Microbiology 71(11), 7107-7112. Whiley, H., Giglio, S. and Bentham, R. (2015a) Opportunistic pathogens Mycobacterium avium complex (MAC) and Legionella spp. colonise model shower. Pathogens 4(3), 590-598. Whiley, H., Keegan, A., Fallowfield, H. and Bentham, R. (2014a) Detection of Legionella, L. pneumophila and Mycobacterium avium complex (MAC) along potable water distribution pipelines. International journal of environmental research and public health 11(7), 7393-7405. Whiley, H., Keegan, A., Fallowfield, H. and Bentham, R. (2015b) The presence of opportunistic pathogens, Legionella spp., L. pneumophila and Mycobacterium avium complex, in South Australian reuse water distribution pipelines. Journal of water and health 13(2), 553-561. Whiley, H., Keegan, A., Fallowfield, H. and Ross, K. (2014b) Uncertainties associated with assessing the public health risk from Legionella. Frontiers in microbiology 5. Whiley, H., Keegan, A., Giglio, S. and Bentham, R. (2012) Mycobacterium avium complex–the role of potable water in disease transmission. Journal of applied microbiology 113(2), 223-232. 417 Whiley, H. and Taylor, M. (2014) Legionella detection by culture and qPCR: comparing apples and oranges. Critical reviews in microbiology (0), 1-10. White, K.H., Larry Soward, and Glenn Shankle (2007) Harvesting, Storing, and Treating Rainwater for Domestic Indoor Use. Texas Commission on Environmental Quality (ed), Austin. Whittington, R., Marsh, I., Taylor, P., Marshall, D., Taragel, C. and Reddacliff, L. (2003) Isolation of Mycobacterium avium subsp paratuberculosis from environmental samples collected from farms before and after destocking sheep with paratuberculosis. Australian veterinary journal 81(9), 559-563. WHO (2004) Guidelines for drinking water quality, World Health Organization, Geneva, Switzerland. Whon, T.W., Kim, M.S., Roh, S.W., Shin, N.R., Lee, H.W. and Bae, J.W. (2012) Metagenomic characterization of airborne viral DNA diversity in the near-surface atmosphere. J Virol 86(15), 8221-8231. Wilton, S. and Cousins, D. (1992) Detection and identification of multiple mycobacterial pathogens by DNA amplification in a single tube. Genome Research 1(4), 269-273. Winthrop, K., Rivera, A., Engelmann, F., Rose, S., Lewis, A., Ku, J., Bermudez, L. and Messaoudi, I. (2016) A Rhesus Macaque Model of Pulmonary Nontuberculous Mycobacterial Disease. American Journal of Respiratory Cell and Molecular Biology 54(2), 170-176. Winthrop, K.L., Varley, C.D., Ory, J., Cassidy, P.M. and Hedberg, K. (2011) Pulmonary disease associated with nontuberculous mycobacteria, Oregon, USA. Emerging infectious diseases 17(9), 1759-1762. Witty, L.A., Tapson, V.F. and Piantadosi, C.A. (1994) Isolation of mycobacteria in patients with pulmonary alveolar proteinosis. Medicine 73(2), 103-109. Wolinsky, E. (1979) Nontuberculous Mycobacteria and Associated Diseases American Review of Respiratory Disease 119(1), 107-159. Wolinsky, E. and Rynearson, T. (1968) Mycobacteria in soil and their relation to disease- associated strains. The American review of respiratory disease 97(6), 1032-1037. Wolinsky, E. and Schaefer, W.B. (1973) Proposed numbering scheme for mycobacterial serotypes by agglutination. International journal of systematic and evolutionary microbiology 23(2), 182-183. 418 World Health Organization (2006) Guidelines for safe recreational water environments: Volume 2, swimming pools and similar environments. Yajko, D., Chin, D., Gonzalez, P., Nassos, P., Hopewell, P., Reingold, A., Horsburgh Jr, C., Yakrus, M., Ostroff, S. and Hadley, W. (1995) Mycobacterium avium complex in water, food, and soil samples collected from the environment of HIV-infected individuals. Journal of acquired immune deficiency syndromes and human retrovirology: official publication of the International Retrovirology Association 9(2), 176-182. Yajko, D., Nassos, P., Sanders, C., Gonzalez, P., Reingold, A., Horsburgh, C., Hopewell, P., Chin, D. and Hadley, W. (1993) Comparison of four decontamination methods for recovery of Mycobacterium avium complex from stools. Journal of clinical microbiology 31(2), 302-306. Yangco, B.G., Lackman-Smith, C., Espinoza, C.G., Solomon, D.A. and Deresinski, S.C. (1989) The hamster model of chronic Mycobacterium avium complex infection. Journal of Infectious Diseases 159(3), 556-561. Yarom, R., Sheinman, R. and Armon, R. (2010) Legionella pneumophila serogroup 3 prevalence in drinking water survey in Israel (2003–2007). Water Science and Technology: Water Supply 10(5), 746-752. Yaziz, M.I., Gunting, H., Sapari, N. and Ghazali, A. (1989) Variations in rainwater quality from roof catchments. Water Res 23(6), 761-765. Yoder, J., Roberts, V., Craun, G., Hill, V., Hicks, L., Alexander, N., Radke, V., Calderon, R., Hlavsa, M. and Beach, M. (2008a) Surveillance for waterborne disease and outbreaks associated with drinking water and water not intended for drinking--United States, 2005-2006. Morbidity and mortality weekly report. Surveillance summaries (Washington, DC: 2002) 57(9), 39. Yoder, J.S., Hlavsa, M.C., Craun, G.F., Hill, V., Roberts, V., Yu, P.A., Hicks, L.A., Alexander, N.T., Calderon, R.L. and Roy, S.L. (2008b) Surveillance for waterborne disease and outbreaks associated with recreational water use and other aquatic facility-associated health events—United States, 2005–2006. MMWR Surveill Summ 57(9), 1-29. Young, L.S. and Inderlied, C.B. (1990) Special Report: Mycobacterium avium Complex Infections: A Comprehensive Overview of Diagnosis and Treatment. AIDS Patient Care 4(5), 10- 18. Zhang, M., Liu, W., Nie, X., Li, C., Gu, J. and Zhang, C. (2012) Molecular analysis of bacterial communities in biofilms of a drinking water clearwell. Microbes and Environments 27(4), 443- 448. 419 Zhou, Y., Benson, J.M., Irvin, C., Irshad, H. and Cheng, Y.-S. (2007) Particle size distribution and inhalation dose of shower water under selected operating conditions. Inhalation toxicology 19(4), 333-342. 420 11. Vita Kerry A. Hamilton was born and raised in Rockville Centre, New York. She graduated high school in 2004 and studied public health and Spanish at Johns Hopkins University from 2004-2008 and was a goalkeeper for the Johns Hopkins women’s soccer team for all four years. During this time she also completed multiple research internships in toxicology (Stony Brook University), chemistry (University of Florida/ University of Campinas, Brazil and Hampton University/ University of Dar es Salaam, Tanzania). Through an accelerated BA/MHS program, she completed a Master’s degree in Environmental & Occupational Hygiene at the Johns Hopkins Bloomberg School of Public health in 2009 and received a certificate degree in Risk Sciences & Public Policy. Kerry was an Association of Schools of Public Health fellow at the U.S. Environmental Protection Agency from 2009- 2011 in the Office of Research & Development in Washington, DC. She was inspired by a mentor to pursue engineering and came to Drexel University to begin a Ph.D. in environmental engineering in fall of 2011. She passed the fundamentals of engineering exam in 2013. Under the guidance of Dr. Charles N. Haas, she has served as both a research and teaching assistant and pursued projects across multiple areas of water quality and risk assessment. She looks forward to applying the many skills she has learned at Drexel to a new set of challenges. 421 Drexel Publications: Hamilton, K.A., Ahmed, W., Palmer, A., Smith, K., Toze, S., Haas, C.N. (2017) A seasonal assessment of opportunistic premise plumbing pathogens in roof-harvested rainwater tanks. Environmental Science & Technology, in press. LeChevallier, M.W., Bukhari, Z., Jjemba, P., Johnson, W., Haas, C.N. and Hamilton, K.A. (2017) Development of a risk management strategy for Legionella in recycled water systems (WRF12-05), WateReuse Research Foundation, Alexandria, VAHamilton KA, Weir MH, Haas CN. In press. Hamilton, K.A., Weir, M.H. and Haas, C.N. Dose response models and a quantitative microbial risk assessment framework for the Mycobacterium avium complex that account for recent developments in molecular biology, taxonomy, and epidemiology. Water Research, 2017, 109, 310-326. Ahmed WA, Hamilton KA, Vieritz A, Powell D, Goonetilleke A, Hamilton MT, Gardner T. Microbial risk from source-separated urine used as liquid fertilizer in sub- tropical Australia. Microbial Risk Analysis, 2015. In press. Ahmed W., Staley C, Hamilton KA, Beale DJ, Sadowsky MJ, Toze S, Haas, CN. Amplicon-based taxonomic characterization of bacteria in urban and peri-urban roof- harvested rainwater stored in tanks. Science of the Total Environment, 2016, 576, 326- 334. Hamilton K, Ahmed WA, Sidhu J, Palmer A, Hodgers L, Haas CN, Toze S. Public health Implications of Acanthamoeba and multiple opportunistic pathogens in roof- harvested rainwater. Environmental Research, 2016, 150:32-327. Ahmed W, Hamilton K, Gyawali P, Toze S, Haas CN. Evidence of avian and possum fecal contamination in rainwater tanks as determined by microbial source tracking approaches. Applied Environmental Microbiology, 2016, 82:14, 4379-4386. Hamilton K, Haas CN. Critical review of mathematical approaches for quantitative microbial risk assessment (QMRA) of Legionella in engineered water systems: research gaps and a new framework, Environmental Science: Water Research & Technology, 2016, 2:599-613. Prasad B, Hamilton K, Haas CN. Incorporating time-dose-response in Legionella outbreak models. Risk Analysis, 2016, DOI: 0.1111/risa.12630. Ahmed WA, Harwood V, Nguyen K, Young S, Hamilton K, Toze S. Utility of Helicobacter spp. associated GFD markers for detecting avian fecal pollution in environmental waters of two continents. Water Research, 2016, 88: 613-622. 422 Ryan M, Hamilton K, Hamilton M. Evaluating the potential for a H. pylori maximum contaminant level. Risk Analysis, 2014, 34(9):1651-1662. 423 Manuscripts under review: Hamilton KA, Ahmed WA, Haas CN. Human health risks due to Legionella and Mycobacterium avium complex (MAC) from potable and non-potable uses of roof- harvested rainwater. Schwarz K, Hamilton KA, Toze S, Sidhu J, Pritchard D, Li Y. The effect of harvesting on the generation of aerosols containing enteric microorganisms from biosolids-amended soils. Submitted to Environmental Research. Hamilton KA, Ahmed W. Pathogens in stormwater: Health risks and implications for reuse. Elsevier book chapter. Pu, G, Hamilton KA, Stillwell C, Montalto F, Haas CN. Reducing urban runoff on ultra- urban land: the ability of a lined stormwater treatment wetland for attenuating runoff and pollutants. Submitted to J. Sustainable Water in the Built Environment. Manuscripts in preparation: Hamilton KA, Parrish K, Ahmed W, Haas CN. Water quality assessment of rainwater barrels in Philadelphia, US. Hamilton KA, Digiovanni K, Rakestraw E, Montalto F, Haas CN. Systematic review of contaminant occurrence in roof-harvested rainwater from North America and Europe and health risk prioritization. ULOQ 8.66E+03 76 1.12E+05
ULOQ >ULOQ 101 2.00E+05